Page 1




Can smart devices really understand us? p. 30

AI puts precision back into medical treatment p. 12

Investigating the secrets of neural networks

One atom at a time

A revolutionary quantum state

Scenes from Flow Festival. See page 50. Photos: Anni Kääriä, Mikko Raskinen

contents This issue is about artificial intelligence.

12 Precision therapy

with AI and a digital twin for everyone?

30 How does your smart

helper function? A brief glossary of AI terms.

40 Curious AI and

an unusually intelligent workmate.

Jaakko Kahilaniemi

5 Openings – Samuel Kaski and FCAI. 6 Now – Minor news, major issues. 10 Wow – Heli Salomaa is an artisan of the video game industry. 11 Oops – Jyri Hämäläinen’s leap into the world of business. 12 Theme – Precision therapy with the aid of AI. 18 Theme – Kasperi Mäki-Reinikka collides art and technology. 20 Who – Miika Aittala explores neural networks. 24 On science – Atomic layer deposition can work wonders. 26 On science – Tuomo Suntola awarded the Millennium Technology Prize. 27 Cooperation – Espoo aims to be a leading smart city. 28 On science – Materials research one atom at a time. 30 Theme – Can smart devices really understand us? 36 On science – Biomaterials challenging plastics. 39 Campus – Väre art collection opens to the public. 40 Entrepreneurship – Mathias Berglund is working on artificial general intelligence. 42 On science – Yu Xiao developing an automatic tacit knowledge engine. 43 Wow – Ilari Niitamo wins Apple Design Award. 44 Current affairs – Learning to learn and other trends. 46 Doctoral theses – Otto Mustonen and a revolutionary quantum state; Julia Talvitie and microlitter; Eeva-Lotta Apajalahti and the energy system. 58 Everyday choices – Petri Myllymäki does not see AI as a threat. 50 Cooperation – Flow Festival.

Photographer Jaakko Kahilaniemi shot several of this issue’s portraits and illustrated the main article. In September 2018, he won the €10 000 ING Unseen Talent Award for his series Nature Like Capital, which explores humanity’s complicated relationship with nature and climate change.


Veera Konsti

On the job


technologies don’t change the world on their own. They’re entangled in the ideas and stories we act out to make sense of the world. If the story of artificial intelligence has, all of a sudden, become one of a revolution to come, we need to untangle the knots that have, up to now, kept the things around us together and ticking. At Aalto, I’ve mainly worked on AI-related themes, and my own epiphany has been to see that, what­ ever will become of the revolution, everything that AI touches along the way will come undone and taken to pieces. Those who reassemble the puzzle will have the power to change the world. Tapio Reinekoski

WHEN THE STATE of the planet can give

you palpitations, it’s wonderful to see glimmers of light in your work: revolutions born in the forest and solar cells that can capture even the faintest winter sunlight. As a science reporter, I often get to experience minor eureka moments. After interviewing Orlando Rojas, I looked at a garden pine tree and wondered at the material, shaped by evolution, that can withstand almost impossible conditions like Finnish winters, storm winds and pests. I want our readers to share in the wisdom of such researchers. Thankfully, I’m not afraid to ask stupid questions and am proud to wave the banner of understandable standard language. Minna Hölttä

WHILE LAYING OUT the magazine and

reading its articles, I arrived at the idea that actual artificial intelligence is still no more than a theory. For humanity to, in a sense, give birth to AI would represent as significant a milestone as the taming of fire. We still don’t understand what it is that makes intelligence work. I believe the answer will be found in humanity’s own intelligence, in the understanding and, finally, cloning of it. As a graphic designer, I take a positive view of AI. I do not, however, believe that AI can master graphic design for a long time because it is incapable of creativity. Perhaps real AI will arise once we understand where and how creativity emerges from. Wille Valkeisenmäki

PUBLISHER Aalto University, Communications EDITOR-IN-CHIEF Jaakko Salavuo MANAGING EDITOR Paula Haikarainen AD/PHOTO EDITOR Liisa Seppo, Otavamedia OMA Oy GRAPHIC DESIGN IN THIS ISSUE Wille Valkeisenmäki




Venla Helenius, Riikka Hopiavaara, Minna Hölttä, Iiro Immonen, Katrina Jurva, Jaakko Kahilaniemi, Kalle Kataila, Krista Kinnunen, Veera Konsti, Anni Kääriä, Lasse Lecklin, David Lewis, Lucas Millheim, Aleksi Poutanen, Marjukka Puolakka, Erkki Pöytäniemi, Mikko Raskinen, Tapio Reinekoski, Guillaume Roujas, Panu Räty, Kaisa Salminen, Laura Siira, Noora Stapleton, Eeva Suorlahti, Tiina Toivola, Wille Valkeisenmäki, Karoliina Vuorenmäki, Ida-Maria Wikström TRANSLATION Ned Kelly Coogan ADDRESS PO Box 18 000, FI-00076 Aalto TELEPHONE +358 9 470 01 ONLINE aalto.fi, aalto.fi/magazine EMAIL magazine@aalto.fi CHANGE OF ADDRESS alumni@aalto.fi PRINTING COMISSIONED BY Unigrafia Oy PREPRESS Aste Helsinki Oy PRINTING Grano Oy, 2018 PAPER Maxioffset 250 g/m2 (covers) & 120 g/m2 (pages) PRINT RUN 5 000 (English edition), 30 000 (Finnish edition) SOURCE OF ADDRESSES Aalto University CRM Partnership and alumni data management PRIVACY NOTICES aalto.fi/fi/about/contact/services/it/ tietosuojailmoitukset ISSN 2489-6772 print ISSN 2489-6780 online


COVER Ida-Maria Wikström CONTRIBUTORS IN THIS ISSUE Jan Ahlstedt, Tiina Aulanko-Jokirinne, Anna Berg, Jon Grönvall, Anni Hanén,


LJÖ M Ä RKT 4041 0955 Painotuote


AI is

far from ready

is a good business at the moment, and ­ I would pitch a career as a professional AI commentator to students, if a handsome number of them did not exist already. Don’t take me wrong – since AI will affect all of us, as citizens we all should participate in discussing how to use AIs for good and not be evil. And it is refreshing and even important to speculate on scifi scenarios. It would just be nice to hear of something more imaginative than the terminators for a change. In July I gave an interview to Sveriges Radio, and even though it was mostly about something completely different, the headline was along the lines of “will AI ultimately take over”. AI is not complete yet, in fact, we only have very simple AIs so far. To get the benefits AI promises, someone has to do the hard work of basic research needed for discovering the new principles and developing the AIs. And someone has to understand the AIs, what is possible with them, and what kinds of changes can their use have in work, society, environment, and ourselves. To a large extent, that someone is us at the universities. At Aalto, we have rolled up our sleeves and taken as our mission to develop the Real AI, which is the next step from the current simple AIs; it should be able to work with us real humans in the real world. Even though Aalto has both long traditions and excellent research in the field at the moment, this goal we obviously cannot reach alone. We have recently launched the Finnish Center for Artificial Intelligence FCAI together with the University of Helsinki and VTT Technical Research Center of Finland, partnering with multiple other top-level international research centers. The best teams from both Aalto

and our collaborators, including but not restricted to the three initiator institutes, are both welcome and needed in this mission. Many of the currently most active contributors are focusing on machine learning, which makes a lot of sense as that is where Aalto is particularly strong, but the mission is fundamentally interdisciplinary and the most interesting stuff happens in between the traditional academic silos – a large number of groups in other fields is already involved and more are welcome! Even the current simple AIs are highly effective, and we have seen nothing yet. Some pioneer companies have already realized that fundamental solutions in AI give fundamental competitive advantages also in other fields, and also noticed that deep AI expertise is the bottleneck in both AI and AI-driven innovations in other fields. As a result, they are already working with university research groups towards the next breakthroughs. FCAI welcomes company partners interested in working in this mode, to solve important problems together. We already have

Kalle Kataila

TALKING ABOUT artificial intelligence

a number of partners, but welcome particularly additional initiatives to together solve big fundamental challenges, from which opportunities will then follow for both parties. It does not hurt that the big interesting problems will also attract the very best students. Artificial intelligence will shake also education and universities to the core. To prevent the cobbler’s children from going barefoot, I have a challenge for Aalto and other universities. Soon it will not suffice that universities only provide the facilities for intelligent people to work with. We need to provide a healthy dose of intelligence – artificial at least – to the way universities operate as a whole. Samuel Kaski Academy Professor Department of Computer Science AALTO UNIVERSITY MAGAZINE 23 \ 5


Mikko Raskinen

School of Business will move to Otaniemi

The new building has four storeys. Its ground floor will contain a restaurant, cafe and teaching premises, while the higher floors will house teaching, meeting and group work facilities as well as office space. THE SCHOOL WILL move from Töölö,

Helsinki, to Otaniemi in Espoo on 15 February 2019. The new building will be complete by the end of 2018 and is located at the heart of the campus. Next door, the School of Arts, Design and Architecture started operating in its new facilities in September 2018. Aalto University metro station and the new shopping centre A Bloc are in the vicinity. The buildings were designed by Verstas Architects. The faculty and staff of the School of Business will move into the new building,

and the master’s and doctoral teaching will start there from 25 February 2019 onwards. Undergraduate teaching will continue as at present in the Otaniemi Undergraduate Centre and in Mikkeli. Although it is new, the four-storey building exudes the famous spirit of the School of Business and its interior design gets some of its inspiration from the unique solutions used in Töölö. In all stages of the new building’s design, the users have been consulted – both staff and students, and alumni as well. The move to Aalto’s shared campus

opens up truly new opportunities for other Aalto staff and students. Multi­ disciplinary studies and research are made easier when everybody is close to each other and the people get to know each other – meeting others will be easy. The current main building of the School of Business in Töölö will continue to be owned by Aalto University. After the move, the building will be renovated. The Arkadia and Chydenia buildings on the Töölö campus have been sold and will be used for other purposes once the School of Business has moved out.

BiographySampo displays the lives of renowned Finns THE BIOGRAPHYSAMPO is an open data

service that enables anyone to easily browse and study the information in biographies and related data sources. The application is based on more than 13 000 short biographies in the National Biography of the Finnish Literature Society and other databases. The material has been enriched with information from other sources: for instance, the collections of the National Library, National Archives, Ateneum Art Museum’s collections, and Wikipedia. The service compiles and links infor6 / AALTO UNIVERSITY MAGAZINE 23

mation about the lives and networks of renowned historical Finns into a new kind of knowledge graph on the semantic web. “With language technology, the texts have been transformed into a data service that can be used to study Finnish history and make connections between people, places and events in surprising ways,” says the project’s director, Aalto University Professor Eero Hyvönen. The BiographySampo can also be used to study groups of people and social networks based on time, place, professional

activities, or various topics. In the biographies of male Members of Parliament (MP) in Finland, for example, the word ‘lead’ is mentioned much more often than in the biographies of female MPs, where the word ‘family’ features heavily. The BiographySampo is a new edition in the Sampo series of linked open data applications developed by Aalto University and the University of Helsinki. The applications already released include the CultureSampo, BookSampo, and WarSampo. The award-winning applications have hundreds of thousands of users.

Hate speech-detecting AIs are fools for ‘love’ HATEFUL TEXT and comments are an

Wille Valkeisenmäki

ever-increasing problem in online environments, yet addressing the rampant issue relies on being able to identify toxic content. A new study by the Aalto Uni-

versity Secure Systems research group has discovered weaknesses in many machine learning detectors currently used to recognize and keep hate speech at bay. Bad grammar and awkward spelling – intentional or not – might make toxic social media comments harder for AI detectors to spot. The team led by Professor N. Asokan put seven state-of-the-art hate speech detectors to the test. All of them failed. Modern natural language processing techniques (NLP) can classify text based on individual characters, words or sentences. When faced with textual data that differs from that used in their training, they begin to fumble. ”We inserted typos, changed word boundaries or added neutral words to the original hate speech. Removing spaces between words was the most

powerful attack, and a combination of these methods was effective even against Google’s comment-ranking system Perspective,” says Tommi Gröndahl, a doctoral student at Aalto University. Google Perspective can be fooled by introducing simple typos, such as removing spaces or adding innocuous words like ‘love’. A sentence like ‘I hate you’ slipped through the sieve and became non-hateful when modified into ‘Ihateyou love’. The researchers note that in different contexts the same utterance can be regarded either as hateful or merely offensive. Hate speech is subjective and context-specific, which renders text analysis techniques insufficient as standalone solutions. The study was carried out in collaboration with researchers from University of Padua in Italy.

guage first developed in the 1960s,” says Professor Jussi Rintanen. The innovation developed by Rintanen and his colleagues automates these kinds of extensive projects: the need for conventional programming will decrease and programming will get considerably easier whilst the development costs go down. In conventional programme development, a programmer focuses on the details of the code, while in the new technology the programmes are synthesised by search methods and logical inference from a high-level abstract specification. “The goal is to make software more flexible and to get it to better understand

how the world outside the information system operates. For instance, in health care, information systems could take more responsibility in performing administrative tasks and decision making.” The threshold for trying out all kinds of business models and ideas will also lower as the production of software becomes cheaper and faster. For instance, the costs of establishing and designing websites and online stores will fall and they can easily be made more versatile than now. Business Finland has granted €678 000 for the commercialisation of this project.

11 901 applicants Ida-Maria W ikström

applied to Aalto University in the 2018 joint application for higher education. The number increased by 19% from 2017. There were altogether 1 641 study places available so the intake was approx. 14%.

Rid of routine coding LARGE INFORMATION systems projects

suffer from the same basic problem: a massive amount of routine programming work whose management is extremely difficult. Building entire systems from scratch would require a vast amount of work. This is why old systems often end up being expanded and modified, even though the original system or the end result might not meet the needs. “Large information systems are often based on ancient program code and old-fashioned programming languages. For example, the Apotti system used by the healthcare sector in Finland is built on old program code in MUMPS, a lan-


PRO NEMUS is Metsä Group’s visitor

AN OUTFIT designed by Juha Vehmaanperä

was showcased alongside collections by more than twenty young Aalto University designers at the Näytös 18 fashion show.

Metsä Group

Guillaume Roujas

centre in Äänekoski. Its lobby is adorned by Uusi kulottuvuus (Controlled Fire), an artwork designed by Aalto University students. The seven-metre installation is made out of pulp and inspired by smoke rising from a controlled burn in the forest. It was created by Karoliina Heikkinen, Tomi Jeskanen, Annamiia Suominen and Daniela Weintraub.

SEVEN SHORT FILMS by members of

Jon Grönvall

the Aalto community were screened by the Helsinki International Film Festival Love & Anarchy’s L&A Shorts competition. Ilja Rautsi’s fictional Helsinki Mansplaining Massacre is a horror comedy about a woman who, after a car crash, winds up at a Christmas party thrown by a group of men.



David Lewis

Aalto University Shop serves customers in the new Väre building on Otaniemi campus. On offer are books, clothes, gifts, art and handy items like umbrellas to fend off the autumn rain.

decoration fair’s Talentshop 2018 event presented four of the most interesting Finnish designers of the moment, including Hanna-Kaisa Korolainen and Hanna Särökaari from Aalto. Korolainen’s exhibition The House of Love brings old furniture together with Finnish ryijy tapestries and ceramics.

Eeva Suorlahti



Wow! Heli Salomaa at Remedy’s scanning studio, where the costumes of virtual game characters are modeled in digital format.

Text: Tiina Toivola Photo: Jaakko Kahilaniemi

Artisan of the gaming world This costume designer switched to virtual clothes. “I’VE BEEN A GAMER since I was ten.

Even back then, I’d wonder who designed the costumes worn by game characters. The video game industry felt distant, however, as I prefer to make things with my hands,” says Heli Salomaa. She recently graduated from the costume design Master’s programme at Aalto University. “I had worked on theatre and performance costumes prior to my studies, but as a person I crave the new and wanted to test the boundaries of costume design. My studies introduced me to laser cutters and 3D printers at the University’s Fablab, so I began to integrate these methods into my costume designs. At Fablab I realised that technology isn’t so hard after all. The video game industry no longer felt like a utopian dream, it was a potential workplace instead.” After this insight, Heli Salomaa freshened up her CV. A couple of years ago, she took part in the Aalto in the Game fair, and then things started happening rapidly. She met a recruiter from Remedy Entertainment, a maker of realistic big-budget AAA video games, at the fair and spoke about her interest in getting involved in game character design. Her timing was impeccable; Remedy’s previous game, Quantum Break, had just been released and a new game called Control, which was scheduled for release 10 / AALTO UNIVERSITY MAGAZINE 23

in 2019, was in its early design stages. Remedy’s lead character artist had already suggested getting help for designing the clothing of the game’s characters. Salomaa was directed to speak with him, and started working as a costume artist a few months later.

Thesis earns distinction

Making games that imitate the real world is a complex process that requires seamless teamwork. During her first year, Salomaa got to know in detail what the members of the design team were doing and what stages she could get involved in. Salomaa combined the linear process of physical costume design and Remedy’s character design process to form a new virtual clothing production line, creating something that has not been employed elsewhere. The makers of international smash hits Grand Theft Auto 5 and L.A. Noire, for example, had hired big-name costume designers from outside the industry. Their approach to the game characters resembled the costuming of a TV show, i.e. they employed analogue methods. “The people at Remedy are incredibly open-minded – they hired me even though my competence was based on analogue costume design. We didn’t even share a vocabulary at first. I also learned

how to use the necessary software, starting from Photoshop, while working at the company,” Salomaa says. “Game makers create huge worlds and technology is taking big leaps forward. The realistic digitisation of physicality, such as architecture or the interaction between the human body and clothing, creates new types of interfaces for association. I’ve been involved with the game Control almost from the very start.” Salomaa wrote her thesis on the process of virtual character design and received two awards for her outstanding work. Her distinguished thesis documented the digital costume design methods and programs employed by Salomaa that any costume designer wishing to find employment in, for example, the video game industry or animation, must master. “Anything’s possible in the virtual world. I can design things that cannot be realised on stage, such as costumes that defy the laws of physics.”•

Costumes designed by Heli Salomaa will be seen in the following productions in 2019 • music and dance film Mirages. Composer: Kaija Saariaho; screenplay and direction: Marikki Hakola (mirages.fi) • video game Control. Producer: Remedy Entertainment (controlgame.com)


Academic calendar

sets the pace Dean, Professor Jyri Hämäläinen’s leap over to the business side took some time.

Text: Paula Haikarainen Photo: Iiro Immonen Illustration: Ida-Maria Wikström “I STUDIED MATHEMATICS at the University of Oulu

around the turn of the 1990s while also holding a job at the Department of Mathematics. I had just completed my doctorate in and was working as a lecturer when, right before summer, Nokia approached me with a job offer. They seemed to be in a hurry, as the interview was set for the very next day. Actually, it wasn’t really a job interview, more like a convivial discussion. We soon reached a shared understanding on the job functions. I went to speak with the head of the maths department the next day, and told him I was considering a move to the corporate world. However, extensive plans for teaching in the coming academic year had already been made, and the teaching of optimisation was largely set to be performed by me alone. The department head and I came to the conclusion that the best time for me to leave would be the following summer. I understood that a one-year delay was quite long, but, having grown up in an academic environment, I considered it natural. I was thus surprised when my proposal was met with a drawn-out and awkward silence at our employment terms negotiation. They were expecting me to show up for work full of vim and vigour the next week. I noted that I hadn’t really looked at the calendar when considering the matter. And, as I’d made some commitments, I felt obligated to fulfil them. I’d be ready to start one year from now. This didn’t really elicit any whoops of joy at the company, which was powering ahead at the time, but we did reach an agreement – and a year later, I joined Nokia in the capacity of a product development technical expert. When I started, I was, to my knowledge, the only doctor amongst hundreds of engineers at the unit attached to a base station factory, and my doctorate was in mathematics to boot. I became sort of a one-man problem solving office to whom the engineers would come with their maths-related problems. This turned out to be a super-efficient way to learn about the industry in practice. I became so interested in the telecommunications industry that I completed a doctorate in that field as well. Two different worlds collided at that junction of my career. I’d become used to the traditional university

environment at the maths department, which did not engage in any external ventures, industrial cooperation or projects back then. I lived in the world of academia, observing the academic calendar. Let’s just say that one week of the corporate agenda corresponds with at least half a year in the traditional academic calendar. My transition to industry was a rewarding experience, which lasted a total of nine years, steering my life path in a fresh direction. Since the beginning of 2008, I’ve served as a Professor at Aalto and, naturally, my field is telecommunications and mobile networks. But mathematics remains my secret mistress to whom I return from time to time.”•

MAGAZINE 23 \ 11


AI puts precision back into medical treatment


Text: Tapio Reinekoski Illustration: Jaakko Kahilaniemi

We don’t receive treatment, medication or operations as individuals, but as averages and aggregates. Artificial intelligence could tap into the enormous volumes of data not only on our bodily systems and genetic heritage, but also on various pharmaceutical substances – and help physicians design our treatments individually.





he last time medicine treated people as individuals, our cures were leeches, mercury, and herbal mixtures, while the most fortunate found pain relief from opium. As doctors engaged in trial and error, patients expired. Medical treatments have been founded on scientific evidence for only a couple of hundred years. Over this time, both technology and means of population control have developed so much that effective treatments can be mass produced. But while the majority of people receive adequate care, this shift has meant giving up on treating people as unique individuals. The information needed to treat each and every one of us individually exists, however. Data on a person’s genetic heritage, health status and treatment history, lifestyle and activity are available in abundance. It’s just too much for humans to process.

Algorithms seek out precision vaccines

One person out of a hundred will get a rash from a vaccine. A common pain killer will seriously poison one in ten thousand. Yet both are considered decent medicines, as functionality and harm are always assessed according to average effects. How a new vaccine, antidepressant or chemotherapy drug is absorbed is unknown. Likewise, how it affects and lives in particular person’s body remains open. To get through the data chaos and find answers, we would need to process superhuman volumes of information with even more superhuman speed. We need help. We need more intelligence. For a long time, the University of Helsinki and University of Oslo Professor Jukka Corander has been developing methods and algorithms for statistical inference, which help create vaccines for diseases like diarrhoea, pneumonia and meningitis that afflict the world’s poorest regions. 14 / AALTO UNIVERSITY MAGAZINE 23

Corander’s methods can screen and test different candidates for the next vaccines and pharmaceuticals, and examine their combined effects in certain genetic populations. At the same time, his creations learn constantly and begin to invent new combinations –the kinds of combinations that humans would struggle to come up with and that would be impossible to find through laboratory experiments. “Our AI models are 10 000 times faster than the testing methods currently in use. We can employ them to make reliable forecasts of how bacteria and viruses will react to different vaccines and how long a vaccine will be able to prevent disease. AI essentially functions like a digital assistant that helps drug developers discover entirely new combinations of pharmaceuticals and pathogens.” There’s an enormous amount of variables in bacterial populations, antibodies and the human body for humans to process. That’s why existing vaccines are not optimal: they are not fully effective for everybody nor are they free of the risk of adverse effects. The effectiveness of a vaccine can also wane; a bacteria population evolves because its genes try to retain their ability to multiply. Corander tries to predict these changes with the help of AI, so that vaccines would retain their bite for as long as possible. “We can already simulate enormous amounts of information from different sources: laboratory test results, bacteria and virus genes as well as their transformations during infections, and variations in the human immune system.”

Sensors on your skin and in your pocket

Would you worry less about a grandparent recovering from an operation at home if they had a bracelet, or even a carpet, capable of alerting an ambulance? Are the drugs prescribed for after-operation care sure to work? Does your family carry a gene that exposes you to cancer? In a couple of decades, these concerns may feel



as antiquated as an old-school, mercury thermometer does now. Everyday devices, like smart phones and watches, can already be used to collect data on your state of health. The manner in which this information is actually taken into use and made compatible with the data that flows in our body is a task for AI—and the key challenge in Aalto University Professor Simo Särkkä’s research. “We’re developing models for combining sensor-gathered data with the information available on the human genome, bodily functions or disease properties. There’s a lot of noise and surplus information in all data, so the models need to be taught to extract only the relevant information in a reasonable amount of time, and to combine information coming from different sources,” Särkkä explains. A smart phone can provide data on a person’s movements and activity in many ways. Its magnetometer detects the Earth’s magnetic field, from which your position and movements can be deduced. Information on speed, acceleration and the position of the phone is provided by the accelerometer, while a gyroscope detects spin. The location of the device can be determined with the aid of a wireless network, the 4G network, Bluetooth connections and GPS signals. “For example, pulse can be measured with the motion sensor and gyroscope by placing the phone in your breast pocket. The sensors then detect the movements caused by heartbeat.” Särkkä and his colleagues are studying how to translate a person into parameters that can be used to construct a comprehensive yet simple model of that individual. “A lot of different kinds of measurable data can be obtained on human physiology. ECG graphs, which are measured with several electrodes, provide a good picture of the electrical activity of the heart, for example. Cardiac mechanics can also be modelled at the same time – as can multiple bodily functions. In principle, it 16 / AALTO UNIVERSITY MAGAZINE 23

would be possible to use machine learning methods to automatically create a complete model of a person, but we’re trying to assemble a much lighter model.” Smart phone and watch apps downloaded from an app store are for self-treatment and, at best, might serve as preventative tools. In order to be accepted for clinical use, a monitoring device must undergo years of studies and testing. Särkkä is cooperating with the Hospital District of Helsinki and Uusimaa to develop a wearable measuring device for diagnosing heart disease. “Days spent in hospital and, consequently, treatment costs would reduce, and patients could return home faster to recuperate. ECG analysis takes a lot of time for the physician, and this can be cut through automation. AI could combine different measurement results and make recommendations for the physician to focus on certain parts of the data accumulated on the patient.”

Screening blood on the spot

Data gathering becomes especially difficult and sensitive when sensors need to reach inside the body. Associate Professor Tomi Laurila at Aalto University employs new carbon materials in the development of extremely sensitive sensors intended for measuring, for example, the concentrations of different neurotransmitters in the brain or the spread and effects of pharmaceuticals in the body. “In order for the sensor surfaces to react to just the right substances and ignore the wrong ones, we need to know the surface atom by atom and how each will react to various substances and their combinations in the body. Machine learning enables us to model carbon structures up to 1 000 times larger than with current quantum mechanical methods – and not lose precision,” Laurila says. “Entirely new physical phenomena have been revealed in our sensor materials thanks to AI.”


The effects of pain killers, for example, are highly individual. Even everyday ibuprofen can cause stomach aches and internal bleeding for many people, never mind the harmful effects of powerful opioids. Laurila believes that, in future, pain killer concentrations – and their likely individual adverse effects – could be determined from the blood count without leaving the doctor’s office. AI-utilising analysis tools would screen blood for desired substances and immediately predict the bodily reactions they cause, eliminating the need to wait days or weeks for test results to come back from a lab. “We have trained machine learning methods to identify carbon in particular, but we’ve also succeeded in teaching AI to detect other substances, like oxygen or hydrogen, as well. The biggest challenge for biological and medical measuring is selectivity: how to extract the right signals from all the noise.” Laurila notes that the experimental development of sensors is, a little like old-fashioned medicine, still largely based on trial and error. Combining machine learning and computational methods speeds up and rationalise experimental research considerably—while also tailoring the sensors to directly detect the desired substances, pharmaceuticals, proteins and mediators.

A digital twin for everyone

Even though the biggest upheavals AI will introduce to health care are yet to come, the most ambitious vision is clear.

“The personalisation of medicine will peak when we can create a digital twin for everyone,” says Academy Professor Samuel Kaski of Aalto University. A model would be created for people individually, combining biological data, clinical examination results, sensor data and information on the patient’s lifestyle, environment and bodily system. “We can now simulate the effects of drugs and treatments on the general level. Digital twins would mean we could do so for each individual. You could in practice query the model tailored for you about anything health-related – just ask, what if ?” AI alone can’t and won’t treat anyone. Determining what to examine and what kind of treatments patients need will remain the job of physicians. A digital twin could provide logical recommendations, evaluate the precision and reliability of its results as well as make an understandable case for its findings and rationale. “Twins could also help in disease prevention. The model could, for example, give an assessment based on diet, exercise, various risk factors, and genetic disposition. Paired with a smart device, the twin could encourage people to do things that are beneficial to their health – in way that they would actually like to observe.” • AALTO UNIVERSITY MAGAZINE 23 \ 17


Artificial intelligence

meets art

The story of AI will not be fully explored without the inclusion of art, believes doctoral student Kasperi Mäki-Reinikka, who wants to collide art and technology – while exploiting the tensions that exist between them. Text: Marjukka Puolakka Photo: Veera Konsti A GROUP OF FRIENDS had maintained

close contact, shared ideas and played role-playing video games in their parents basements since primary school. Their meetings became more infrequent once one of them began, after attending the Kallio Upper Secondary School of Performing Arts, to study at the University of Art and Design Helsinki while the others chose paths leading to cognitive science and electrical engineering. “When we’d meet, an argument would always arise on the differences between science and art. Should we search for objective truths or subjective experiences? We could never reach a consensus, so, one evening, I came up with the idea of us creating art together,” says artist, visual arts educator and doctoral student Kasperi Mäki-Reinikka.


And thus the Brains on Art art collective was born. They decided to see where they’d arrive at by replacing dead-end arguments with the making of art using the skills and means available to students of art, cognitive science and computer science. Their first exhibition piece was completed in 2013. At the same time, Mäki-Reinikka’s master’s thesis sketched out the negotiative and contrast-seeking method of working created in the collective. Studies in the Art and Technology minor programme in the newly-founded Aalto University only spurred him on. “Our artistic activities are entirely based on the friction between different fields. In spite of the tension, or perhaps thanks to it, we create works that are both artistically interesting as well

as scientifically and technologically challenging.” Brains on Art has created poems based on recordings of the electroencephalographic patterns of viewers as well as a performance art piece where a performer sways from side to side in tune with the fluctuations of the Helsinki Stock Exchange.

Reaching across fields

Mäki-Reinikka is currently a doctoral student at the Department of Art. His research explores fresh working practices that connect art and science. “Putting it bluntly, I’m studying how tech students and art students put in the same room can succeed in creating art together. I’m examining what they should know about each other’s fields

and what kind of debate would enable the creation of new and artistically interesting works.” In addition to the perspective of a media artist, Mäki-Reinikka’s research takes advantage of his teaching work in the University-wide Art Studies programme, which encourages students in tech, business and art to engage in shared art projects. As art content coordinator of Aalto University Junior, he steers schoolchildren towards phenomenonbased learning. Mäki-Reinikka also participates as an artist member in the advisory working group of the Finnish Center for Artificial Intelligence initiated by Aalto University, University of Helsinki and VTT. “AI affects society as a whole. I’m delighted that FCAI wants to introduce the perspectives of all fields, art included, into the technical development of AI.”

New aesthetics through AI

AI is associated with moral and ethical questions that art is especially suited to consider. “Artists have always been quick to adopt new techniques, such as photography, video and the Internet, to employ in novel ways. Art can equally well introduce fresh openings for developing AI and machine learning. Could machines, for example, produce aesthetics of an entirely new kind?” AI is also hard to pin down, MäkiReinikka notes. “Today’s smart phones would have been considered AIs 50 years ago. Be that as it may, the future will always bring forth new instruments for artists to work with.” The most interesting projects are often those that can be approached from several viewpoints. Art can introduce something to the development of AI that would otherwise be missed. There is an increasing need for dialogue across fields. “Dialogue, which spans different worldviews and concepts of knowledge, is not a threat to each party’s own value foundation and understanding of the world. Getting a peek of the other side is in and of itself valuable and necessary.”

Threat to humanity?

The boundary between the biological and the technological is narrowing as bio- and nanotechnology and cognitive science develop. Our sensory world is expanding through virtual and augmented reality as well as via devices creeping under our skin. When humans

and machines begin to form something new, it is pertinent to ask how a machine can understand what constitutes art or an aesthetic experience. “New technologies carry great cultural significance regarding what, in future, it will mean to be human and what is humanity’s place in the world. Considerations of this magnitude cannot be left to one field alone. Art plays an important role in the adoption of new technologies.” Brains on Art is good proof of how collisions between art and technology can give birth to something completely unique. In a multidisciplinary community, no one gets off easy when solving strange problems that might never be encountered alone. Art also provides a free space for experimentation. “In an art project, tech students and engineers get to experiment and play with their own specialised knowledge and skills in an entirely different way than would be possible in an experimental laboratory.”

Machine meets art

Right now, the Brains on Art collective is pondering how machines will meet art. It’s possible that they’ll find that machines are really dumb, incapable of understanding art. Or maybe not. What if machines do become capable of both creating and sensing art? Would this deprive artists of their work – and what’s the need for human viewers, then? The collective is working on an mac­ hine learning image-recognition system, which uses tens of thousands of pictures provided by the Finnish National Gallery’s archive as raw material. What will happen when the machine viewer starts to classify and assign value to the pictures according to its own logic? “The usual approach is to supply similar images from within a specific genre of art, but we instead keep the experiencing machine in a loose leash and feed it with new works at certain intervals. The end result might start veering in unexpected directions.”

Get inspired, experiment and make it real

Look for opportunities for combining art and science. Check out workshops, follow research and propose a project.

Look at art and make it yourself – Aalto offers excellent courses on art-based thinking.

. Challenge, be critical. There’s no need to agree on everything. Interdisciplinary friction can be transformed into a creative force.

Remember: no field is monolithic. Art and science are created by people who come from many different backgrounds and think in many different ways.

Towards the end of the year, the Aalto campus may very well witness a work that introduces us to a new, aesthetic machine experience. The precise manner in which the piece will be presented remains open. “It could be an AI installation, which produces and experiences art in a specific space, or a portable creature you take with you to the gallery to look at art.” •

Kasperi Mäki-Reinikka graduated as a Master of Arts from the Aalto University School of Arts, Design and Architecture in 2015. He also studied visual art at the Beaux-Arts in Paris in 2009–10 and completed Aalto’s Art and Technology academic minor in 2012. The Brains on Art art collective founded by Mäki-Reinikka has arranged four private exhibitions and participated in numerous joint exhibitions since 2013. Mäki-Reinikka serves as coordinator of art content for Aalto University Junior and as artist member for the FCAI Society within the Finnish Centre for Artificial Intelligence, in addition to teaching in the University-wide Art Studies programme. He is also working on a doctorate at the Department of Art.




Investigating the secrets of

neural networks A stroke of luck at a Tallinn market got the teenage Miika Aittala hooked on graphics. Today, he applies AI tools known as neural networks on graphics problems at MIT in Boston. Text: Tiina Aulanko-Jokirinne Photos: Mikko Raskinen



goal is to create visually realistic worlds in virtual environments. He employs neural networks and 3D modelling as his tools, with the outcomes in high demand in, for example, movie productions. This area of research is trending strongly around the world, and competition for the cleverest methods and researchers is intense. Aittala took his first step towards becoming a researcher as a teenager at the end of the 1990s. He found a copy of the 3D Studio MAX professional modelling software at Tallinn’s Mustamäe market. Inspired by the special effects in films like Jurassic Park and Toy Story, Aittala used the software to create 3D animations and images. He might have ended up producing effects for cinema – but research was to win him over.

From VTT and Aalto to the world Aittala began to study technology and, in the aftermath of a summer job, also ended up completing his Master’s thesis on graphics and augmented reality at VTT Technical Research Centre of Finland. His ambitious work led to doc-

toral studies and a place in Professor Jaakko Lehtinen’s research group at Aalto University. “I retained my interest in graphics after completing my Master’s, and it felt natural to start researching it when the University offered me the opportunity. I was quickly working with distinguished and well-known researchers.” Aittala’s doctoral thesis studied the 3D modelling and reproduction of materials to generate realistic computer graphics. Even then, his aim was to create graphics tools that were as simple, elegant and computationally light as possible. Aittala visited a graphics research retreat organised by MIT Professor Frédo Durand in the United States to talk about his research just a little bit before finishing his award-winning doctoral thesis. They agreed on close research cooperation during their meeting. This is how he started his journey to the world’s premiere technology university and its famed graphics laboratory.

Enormous data volumes enable neural networks The limits of traditional methods have been met in many graphics research problems, but neural networks have AALTO UNIVERSITY MAGAZINE 23 \ 21


enabled fresh approaches to them. Although the basic principles of neural networks have been know for a long time, a proper understanding of their potential has only developed in recent years. “Unlike just a decade or two ago, there is an enormous amount of data, for example photos, available on the Internet, and this can be used to teach neural networks. Computing power is likewise on an entirely different order of magnitude than before.” The computations we now perform in a few days would have taken years in the 1990s, and photos snapped with a phone can, at best, be processed in less than a second. Earlier, there was no inkling that good results would be achieved only with the arrival of great computing power. “A large part of current research resembles charting unknown territory by trying various tools for different tasks,” Aittala says. “I assume that, in twenty years, the overall picture will be much clearer, and the ideas now prevailing will look like stumbling, yet important first steps. The field is moving forward so rapidly that what was considered wisdom just a couple of years ago can have been upturned several times already, however.”

Visually believable worlds

Materials modelling and reproduction is utilised especially in various graphics applications like special effects production or augmented and virtual reality. Neural networks are strong in approximate intuitive inference: if, for example, a picture in itself does not contain enough data, it can be complemented with the aid of computation. This creates


visually credible virtual surfaces, which are more significant in gaming or cinematic applications than perfectly realistic reflections. Neural network inference follows along the lines of human intuition, and will also generate an estimate of the material’s nature based on a “fingerprint” formed by visual clues. Each clue, which describes the material’s surface, forms its own map: a relief map shows surface shapes, specular maps detail the shininess of each pixel on a material’s surface, while reflection and base colour maps describe the other properties of the surface. The different components of an image can be directly utilised in rendering, i.e. the creation of new virtual images frame by frame. “A single movie scene can contain a million things. Each detail on the screen must be in order so that nothing out of place sticks out. Viewers are really good at noticing if the graphics are done wrong. This is why every movie production employs such huge numbers of artists. If I can simplify and ease their work, they can concentrate on more essential artistic issues,” Aittala says.

Flexibility of neural networks behind the AI boom Data is accumulated for neural networks to use for learning how to make independent deductions, and the results of this are applied in, for example, the creation of virtual realities. The inference process performed by the network remains a mystery, however. Neural networks are free to form flexible descriptions or results from the input data, but they provide no explanation for the outcome.

Miika Aittala considers this development stage necessary and interesting. “From a researcher’s perspective, things are happening rapidly and in entirely new ways. Neural networks allow us to tackle problems that were difficult to grasp in any way earlier. The operating model as a whole will start to become clearer once neural networks have been utilised for a sufficiently long time.” There is a long way to go before an independently operating and general artificial intelligence comes to be. Existing neural networks are only taught to solve a single, precisely defined problem at a time. A new research problem always requires a new neural network architecture. Aittala’s new project explores how to remove blur from photographs suffering from camera shake. He also participated in a joint study arranged by NVIDIA, Aalto University and MIT that cleaned up grainy images with the aid of neural networks. This study made the news around the world in July 2018. American graphics technology company NVIDIA has licensed and started to commercialise the findings of Aittala’s doctoral research. This could result in programs game makers could use to generate virtual realities. Applications might also end up in the hands of consumers, for example as devices that can create 3D models of any object or surface. “Neural networks help us achieve results that couldn’t even be dreamed of just a few years ago. It remains to be seen whether this field of research will grow more established, or if the current development, which is characterised by ever-increasing enthusiasm and new breakthroughs, will continue.”•



On science

A layer just a few atoms deep can work wonders The atomic layer deposition (ALD) technique revolutionised information technology in the 2000s. Now researchers are using it to make better catalysts, solar cells of record efficiency and hybrid materials that transport medicines and generate electricity from bodily waste heat. Text: Minna Hölttä Illustration: Ida-Maria Wikström “MY RESEARCH GROUP develops black silicon cells, i.e. nanostructured solar cells. The structure enables the cells to catch sunlight or photons also from a very low approach angle, and the cells can thus generate electricity also in the morning, evening and during the darker seasons. In order to efficiently recover the electrons the photons create on the cell surface, the surface needs to be covered with a thin membrane called a passivation layer. As the surface nanostructures are so small, the only way to manufacture a membrane that matches their shape is to use ALD – other methods would just result in a loose membrane lying on top of the structure. It is important for the functioning of a solar cell that the front side’s thin membrane allows all light to pass through it. The material we most commonly use for these membranes is aluminium oxide, i.e. a mixture of aluminium and oxygen, but any material that forms a good interface with silicon will work. We already hold the efficiency world record for nanostructured solar cells, 22.1%, but are constantly researching new materials – we could find something even better, after all. The electron physics research group I head is rooted in the Electron Physics Lab, which was established in 1969 – and happens to be where the inventor of ALD, Tuomo Suntola, completed his doctorate in 1971. In addition to us, Micronova’s clean rooms nowadays house several research groups that utilise ALD in optoelectronics and photonics applications, for example.”

Professor Hele Savin 24 / AALTO UNIVERSITY MAGAZINE 23

“I STUDY the use of of ALD

in catalysts. Catalysts are substances that speed up chemical reactions without being consumed in them. I specialise in heterogeneous catalysts, solid substances with large surface areas: one gram can easily contain 300 square metres. Catalysts are, of course, also made without ALD, and enormous amounts of them are used in industry. The traditional manufacturing processes aren’t that well controlled, however, and the end result can vary a great deal with respect to, for example, the particle size of the catalyst metal. ALD enables us to make catalysts with a great degree of control and repeatability. This makes them more effective in the catalytic process itself, and also simplifies the modelling associated with testing. ALD can also be used to put surfacing on catalysts that prevents the build-up of metal particles. This keeps the catalysts active for longer, even in harsh conditions. The surface reactions that occur in ALD have always fascinated me. Even though the technology was introduced into industrial use as far back as the 1980s, real understanding of how the method functions has only come afterwards. As our scientific understanding grows, we become able to utilise the technology more diversely.” Professor Riikka Puurunen


when the aim is to make high-quality thin membranes out of known materials, like simple metal oxides. I, however, specialise in using it to make more complex and previously unknown materials. For the last seven years, my group and I have concentrated on hybrid materials, which combine inorganic layers with organic layers with the aid of molecular layer deposition (MLD). My goal is to get the best sides of both inorganic and organic materials to come together in one and the same thin membrane.

Hybrid materials are promising as, for example, thermoelectric materials that convert waste heath directly into electricity. Thermoelectric materials should conduct electricity well and heat poorly, which is a conflicting demand for conventional materials like zinc oxide. Adding organic layers to zinc oxide, which conducts electricity well, reduces the material’s thermal conductivity and also increases its flexibility. Flexibility is important if the aim is to use the thermoelectric material in, for example, clothes. We’ve followed a similar idea when making a thin-membrane microbattery, which is based on flexible hybrid membranes.

The materials used for thin membranes are usually amorphous, i.e. there is no crystal structure in the membrane. My dream is to manufacture crystalline hybrid membranes, and in this way create new, exploitable properties. Hybrid materials typically have very open crystalline structures. These contain suitable nano-scale pores, which can selectively absorb various small molecules. Crystalline hybrid membranes could find use in, for example, sensors, air and water purification, gas storage and pharmaceutical delivery.”

Professor Maarit Karppinen AALTO UNIVERSITY MAGAZINE 23 \ 25

On science

Millennium Prizerecognised technology found in every smart phone AALTO UNIVERSITY ALUMNUS, Doctor of

membranes, which are needed in microcircuits and memory components. Suntola’s innovation is a central factor behind famous Moore’s Law remaining valid up to now: microcircuit performance has doubled every two years while prices simultaneously drop and components get smaller. The technology is widely used in mobile phones, computers and other electronics. It is forecast to spread also to health technology applications. Tuomo Suntola was awarded the one million euro Millennium Technology Prize in May 2018. “ALD has made the ownership of information technology more democratic, and in this way improved humanity’s access to information and opportunities for communication,” Academy Professor Päivi Törmä, chair of the international prize committee, said about the decision to honour Suntola’s work. •

Lasse Lecklin

Science (Technology) Tuomo Suntola developed atomic layer deposition technology and the equipment to manufacture thin membranes in the 1970s. The technique involves a gaseous substance or compound being sprayed into a reactor chamber. The surface being covered binds only as much of the substance or compound as is needed to fully cover itself, i.e. at most one atomic or molecular layer. The excess is then rinsed away, after which another substance or compound is released into the chamber. The process is repeated until the membrane reaches the desired thickness. Use of ALD exploded in the 2000s when the electronics industry realised its potential in the manufacture of semiconductor components. The method makes it possible to manufacture extremely thin electricity-insulating or -conducting

Millennium Technology Prize laureate, physicist Tuomo Suntola held an open public lecture on atomic layer deposition at Aalto University. 26 / AALTO UNIVERSITY MAGAZINE 23


City of Espoo

What will change?

Better services for the citizenry with AI The City of Espoo wins an international award and aims to the top of smart cities. Text: Kaisa Salminen THE CITY OF ESPOO is a member of

the Finnish Center for Artificial Intelligence FCAI, which aims to introduce the most advanced AI methods for adoption by companies, public organisations and society. A number of AI trials have already been performed in Espoo. For example, the social welfare and health department’s customer data was recently analysed with the aid of AI to identify 280 factors, which might predict familyspecific future need for contact by the child protection division. “The objective is to develop preventative services with the aid of AI, and in this way avoid stressful events like children being taken into custody and placed into care,” says City of Espoo Data Analysis Consultant Tomas Lehtinen. Lehtinen says the City of Espoo has an exceptionally large amount of advanced information systems, but they remain underutilised from the AI perspective. “AI makes it possible to sift desirable information from an enormous amount of data quite rapidly. However, not all of the City’s information systems are suitable for research use at present. We need to develop our systems to enable the flexible, reliable and safe gathering of information,” Lehtinen stresses. “For us, the most important thing is the quality of the data, and the way in which information is gathered. The City

needs to be careful because, among other things, our data recording practices have changed over the years.”

Communal development work

Elsewhere around the world, AI experiments associated with the Mobilityas-a-Service (MaaS) concept in particular have progressed far. Espoo’s trials related to its social welfare and health care data have, in turn, aroused interest both in Finland and abroad. The City recently placed first in the Intelligent Community Awards 2018 as recognition of its community-oriented intelligent urban development efforts. The theme of the competition was the exploitation of data in the development of humancentered services. “We want to engage the citizenry to an increasing degree. Our goal is to make Espoo the capital of AI,” Tomas Lehtinen envisions. Lehtinen reckons that, in the initial stages, AI will serve Espoo mostly in the capacity of a support intelligence, which eases and facilitates the everyday work of professionals. Later, AI will probably benefit the City in other ways too, for example through the automation of functions and services.

Shared databases

FCAI’s Corporate Liaison Terhi Kajaste says that Espoo’s decision to

Data Analysis Consultant Tomas Lehtinen lists some AI development areas relevant to the everyday lives of companies or citizens: • AI will serve as a “support intelligence” and make the everyday work of professionals easier. • Research and product development will become faster. Espoo is already an innovation platform for companies, and AI will be introduced more strongly into development work in the future. • Bots will enter the everyday lives of the citizenry. For example, a project is about to be launched that will test how chat and telephone bots can support customer service with regard to frequently asked questions. • Municipal services will improve when AI enables improved forecasting of future service needs. • The significance of Mydata, i.e. an individual’s personal data will be emphasised; people will be better able to manage their own data, which will also facilitate service development work.

join the project supports its goal of providing broad-based service to society. “The creators of AI must have a profound understanding of the principles with which AI makes its decisions, otherwise an AI trained using, say, a municipal database might mistakenly draw conclusions on a discriminatory or some other undesirable basis,” Kajaste says. Tomas Lehtinen says the intention is to consider how the new EU General Data Protection Regulation should be interpreted in cooperation with the other cities and AI researchers participating in the FCAI. “In Finland, the personal identity code offers good opportunities for combining information systems, but after the systems have been merged, a person must no longer be identifiable. The use of various AI-based alarm systems also calls for more detailed debate – can they be used, who will operate them and how will the individual be contacted if necessary.” Terhi Kajaste says Espoo’s own databases and those shared by the municipalities of the capital city region as well as the data they contain are extremely interesting to FCAI researchers who are developing next-generation AI methods.• AALTO UNIVERSITY MAGAZINE 23 \ 27

On science

One atom at a time AI is not only changing the way we process data, it’s changing how we carry out research, period. Text: Katrina Jurva Photo: Venla Helenius EVERY MONDAY morning at 9am,

Milica Todorović grabs a bowl of porridge and sits down to talk with computer scientists about how machines learn. Todorović is one of the many researchers from the natural sciences looking to apply the power of AI in her daily work; by keeping up with what experts in machine learning are doing, she hopes to bring some tiny things together in smarter ways. What kinds of tiny things? Think minuscule. Milica Todorović’s work focuses on how devices from mobile phones to solar panels function at the atomistic level, the smallest unit of ordinary matter. Her days revolve around finding the best composite materials to get the job done. With global challenges like adopting green energy, the world needs these kinds of advanced materials to bring about the next wave of technology and energy solutions. “The forefront of materials engineering is really in making complex materials that combine vastly different properties,” explains Todorović. In her case, this typically involves a mix of organics, generally soft materials like cellulose from wood products, and inorganics, like metals. With the seemingly never-ending list of configurations out there, “we need to shake up our ways of thinking to find new solutions,” she says.

Getting to the bottom of things

Optimizing materials for the technologies of the future is not just about tweaking one small thing here or there, it’s about understanding how the materials inside devices interact at the most fundamental level. Picking the right materials is the first step, yet having these in hand doesn’t 28 / AALTO UNIVERSITY MAGAZINE 23

guarantee smooth function. Choices made at the assembly stage can inadvertently create less-than-optimal structural contacts, ultimately affecting how a device performs. You can select the best possible components, but if you don’t combine them in just the right way or at the right temperature, you’ll end up with something less than ideal. When it comes to the inner makings of materials, say, in electronics, the possibilities of different matches are nearly endless. At the same time, the better the match between materials, the stronger the device becomes. The question is how to get there. Conventional thinking in quantum mechanics, the area of physics that looks at the world at the smallest possible scale, has seen this type of calculation as simply too big to get results at the speed in which they are needed. Until now, researchers have relied on applying certain design principles based on knowing AI could speed up the processes of materials research enormously, say Milica Todorović and Patrick Rinke.

what materials traditionally work well together. This makes a lot of sense. With years of study behind them, scientists have a strong sense of the types of composite materials that suit certain applications and are well equipped to find a solid solution from a pot of options. But in a field where the sheer number of combinations exceeds human analytical capacity, nothing is certain. Studies on composite materials tend to be costly in terms of human-hours and computing time. In the end, there is a lot of guessing and chemical intuition involved. “Our chemical intuition can guide us a long way, we’re just never sure if there isn’t some configuration that we hadn’t considered that might be relevant,” Milica Todorović admits.

“This is why we now invest in AI, because it gives us a kind of unbiased view of the material.”

Changing the way we see

Milica Todorović works closely with Patrick Rinke, a professor in applied physics and expert in computational materials science. Having seen the field develop over the last two decades, they see AI as a tool that can help the team do things not only differently, but better. “We’re using AI as a different way to make sense of the natural world,” Patrick Rinke says. “It not only helps us identify patterns that we as humans can’t see, it allows us to process large amounts of data.” While improving the materials design process has always been the goal, until now scientists have worked to simply speed up the steps they have always used. If finding the right material combinations is like getting from point A to B, the way scientists have improved methods has been by jumping in the car and trying to drive faster and faster each time.

Rinke explains that using AI to refine materials is something different: it’s a paradigm change. “Instead of using a car, you’ve completely changed the approach and used a plane to get there.”

Get quick results, leap ahead

So, how well does it work? According to Todorović and Rinke, very well. Replicating a conventional study on the contact between a molecule and a crystal, one of the smallest material interfaces possible, their AI tool was able to come to the same conclusion in just 3% of the computation time. “It’s about 30 times faster and 100 times less pain,” Patrick Rinke chuckles. AI also allows researchers to focus on the main problems. The process of identifying the contact structure is often so intensive that little time is left to consider properties and, by extension, device function. By harnessing AI’s power for the first task, researchers can place more energy into the second stage of development – and, with any luck, get better results. The tool is not, though, without controversy. Traditional methods, regardless of the field, are tried-and-true. It can take time for new approaches to gain traction, even when they bring about quick results. The fact remains that with many AI methods, the answers are quick but the way in which they are achieved is not always clear. “The relation between the answer and the data is still more opaque than we would like it to be. Still, we trace back because we love to know why!” laughs Milica Todorović.

“It’s not enough to get a good answer – we want to know why our new methods are working so well.”

Creative juices required

With the core understanding of AI technology in the hands of computer scientists, getting to the next step requires top-notch communication skills and the ability to translate knowledge from one world to another. Case in point: Knowing that AI has the power to learn the difference between, for example, dogs and cats in photographs, the team set out to develop an AI that learns from the materials they already know. “We knew that if we could find an image representation for our data, we could use the same tool that distinguishes between cats and dogs,” Patrick Rinke shares. Once they had that, they just needed to train the AI to know that one image means this and another image means that. “We did this 50 000 times and the next time it made the correct prediction,” he says. The methods are out there; taking them on just involves a step or two outside the typical way of doing things. “We have to go outside of our community to learn about these new techniques. It’s really important to have meetings where people can cross-pollinate fields,” Milica Todorović emphasizes. “We need to remember that talking to each other creates value.” •




Can smart devices really understand us? Virtual helpers have introduced interactive artificial intelligence into our everyday lives. The next step is for computers to learn how to relate to us as individuals. Text: Panu Räty Illustration: Ida-Maria WikstrÜm







from Sci-Fi literature into our pockets and on to our tables. Amazon’s Alexa, Apple’s Siri and Google Assistant are taking us away from computer and mobile handset screens – and making the spoken word our new user interface. It is already easy to check news, choose music, order a cab and command household smart devices using voice control. But how are such AI-utilising smart helpers in fact able to understand us?

Phase 1: speech recognition

In order to function, smart helpers must always be on. They’ll hibernate and listen to their environment until they recognise a key word uttered within range. Amazon’s virtual assistant, for example, wake up when it hears its name Alexa. An LED ring on the smart speaker turns blue to indicate that it has awoken. Apple’s Siri also functions with the same principle. When it hears the prompt Hey, Siri, it starts recording and uploading the user’s speech to recognition software stored on a cloud service. The digitised speech is first spliced into short bits only fractions of a second in length. “Everything starts with spectrum analysis, i.e. examining the frequencies found there. Patterns, which describe different sounds, are created in the frequency space,” says Associate Professor Mikko Kurimo from Aalto University’s Department of Signal Processing and Acoustics. All material redundant from the perspective of speech recognition, such as the pitch of the speaker’s voice and background sounds, is removed in conjunction with splicing. “In other words, it tries to find patterns that indicate what speech sounds have been uttered,” Kurimo says. Speech recognition is made more difficult by that fact that we speak incoherently, swallow words and use gestures and utterances. The words we speak can also sound alike, as is the case with, for example ate and eight. “These days, speech recognition is more and more often performed with deep neural networks,” Kurimo says. Deep neural networks mimic the way the brain oper32 / AALTO UNIVERSITY MAGAZINE 23

ates and consist of certain types of simple calculators known as artificial neurons. A neural network becomes efficient when interconnected neuron layers communicate with the neurons of the same and the next layer. In addition to statistical sound models, neural network speech recognition search algorithms utilise language models built with the help of extensive text materials. Language models predict the probability a word will occur after another word as well as the likely way in which it will be pronounced. This helps weed out unlikely words to speed up recognition. “A speech recognition application thus performs the task of finding the sentence the user most probably spoke,” Kurimo says.

Phase 2: processing natural language

The aim of natural language processing is to decipher the meaning of text – i.e. identify what the user wants from its digital helper. Neural networks are also utilised in natural language processing. Speech data is scoured automatically for key words and phrases in order to ascertain what the user’s words might possibly relate to. Neural networks are trained for their tasks by feeding them a large volume of data for processing and then comparing their output values to known correct values. Corrections are made until the result no longer improves. After this, the system is capable of operating independently. One project headed by Kurimo has researched the production of automatic descriptions of audiovisual material. Among other things, archived Yle videos were chosen as source material. The developed method is able to simultaneously interpret both the speech recorded on the video as well as the moving video image – and can generate a text description of them. The system was taught by using human-written descriptions of the same videos as points of reference. The size of the databases used to teach deep neural networks is a central factor. This is why commercial digital helpers are being produced by giant corporations like Amazon, Apple, Google and Microsoft. “Major companies have access to extensive databases, and they can perform automation quite easily. It arduous to start making a chatbot from scratch. You have to accumulate a database somehow.”


A brief AI glossary • ALGORITHM Definition of how to perform a task, which opens the details of the process step by step. The term program refers to the expression of an algorithm in a programming language. • ARTIFICIAL INTELLIGENCE Refers to a computer’s ability to engage in intelligent activities. Is a broad and fluid umbrella concept for various inference principles. • APPLIED (NARROW/WEAK) ARTIFICIAL INTELLIGENCE The only type of AI humanity has produced so far. Produces solutions for a predetermined application area by combining analytics and automation. • ARTIFICIAL GENERAL INTELLIGENCE The idea of mimicking human intelligence on computers made popular in science fiction literature. Sometimes an AGI is presented as becoming sentient. Our technology has not even come close to building an AGI for now. • CHATBOT (BOT) A software robot that performs tasks automatically mimicking human interaction. The most sophisticated versions are AI-based collections of programs that can search for spoken or text answers on limited subject matters. • MACHINE LEARNING Methods by which computers can learn to reach a desired outcome independently based on examples and experience. • NEURAL NETWORK A machine learning architecture and algorithms, which have some loose inspiration in how human brains function. They consist of interconnected artificial neurons i.e. simple processors. Adding layers to a neural network deepens it and enables the performance of increasingly complex tasks.



AI IS NOT YET ABLE TO BE A WORTHY PARTNER IN A DEBATE – LIKE IN SCIENCE FICTION MOVIES. frame of reference to which words or gestures refer to.” Research on the interaction between humans and AIemploying systems is nevertheless progressing all the time, and the area of possible application is expanding in tandem. One such application area is using computational models to improve user interfaces, a subject that Oulasvirta’s research group has been studying. For example, a user’s browsing history can be used to reformat a website to a layout that feels immediately familiar to the user. “It is possible to create a more pleasant browsing experience in this way. Headers, for example, could almost always be found in the same spot.” Fresh research subjects have also been discovered within an activity as mundane as inputting text. Coupling cognitive science, a field that researches phenomena related to observation, learning and memory, with AI enables the building of models, which accurately predict how a person’s individual characteristics affect, for example, writing on a smartphone display. When such models are connected with a machine optimiser that simulates alternatives, the user interface can be tailored to suit a specific user. This process has identified smartphone use solutions for older people who suffer from shaky hands, for example. Oulasvirta’s team has also created a new layout for French computer keyboards, which was recently approved by the standardisation authority of France. “All of France will be typing special characters in a way determined with the aid of our optimiser,” he says. A research project dealing with the modeling of emoTowards individualised user interfaces tions is also ongoing. Even the most conversational AI will not, for quite “In the final analysis, the field of AI deals with presome time, be able to serve as a worthy debate partner senting human matters computationally,” Oulasvirta like we’ve seen in so many science fiction movies. says. Professor Antti Oulasvirta from the Department He points out that the familiar journey planner found of Communications and Networking considers it prob- on smartphones is also based on AI, even though most lematic for voice user interfaces that they are unable users would not think of it as an AI application. Oulasto actually understand language. virta’s view on the matter is, however, clear. “AI doesn’t learn language sort of by engaging in phys“Whenever some intellectual capacity can be realised ical and social interaction. It cannot learn the linguistic computationally, it, in my opinion, represents AI.” •

Phase 3: fulfilling the request

The last phase is to fulfil the user’s request. In addition to information retrieved from the net, digital helpers take advantage of, for example, the contact details, location information and calendar on the user’s phone in order to form a better idea of what the user wants. This is why a digital helper can appear surprisingly smart when fulfilling simple requests like connecting a call, looking up weather information or ordering a pizza. But ask a helper like this to tell you what’s going on in Silicon Valley, and it will provide a clumsy answer containing random search results related to the term Silicon Valley. A digital helper would be unable to deduce whether it is being asked about the history, weather or companies active in the area. “They run out of smarts the moment you go beyond their design space,” Mikko Kurimo says. There has also been a shift to employing deep neural networks in generating voices for digital helpers. Speech sounds are always interconnected in natural speech, and incompatible sounds were precisely what made early smart helpers sound so robotic. Today, neural networks perform calculations on the fly to enable the correct pronunciation of the phrases spoken in reply. “A synthetic speech generator is fed the syllables and words to be emphasised as input, and these make the speech sound natural. The generated signal is then transmitted to the user’s terminal device for playback.”


On science

The next material revolution will start in the forest Plastic straws and utensils will soon be history but they will not be missed, as something much better will replace them.

Text: Minna Hölttä Photo: Aleksi Poutanen AT THE START of the year, the EU pub-

lished its first-ever plastics strategy. It is based on the circular economy and related statistics, which indicate the urgency of a change in the culture of material utilization. More than 25 million tonnes of plastic waste is produced in Europe each year. Less than a third is recycled, while the rest ends up as unsorted landfill waste or in the environment. 84% of the waste found on our beaches is plastic, and disposable goods like cotton buds, plates, drinking straws and forks account for 36 / AALTO UNIVERSITY MAGAZINE 23

the majority of this. The EU is now planning a total ban on such products in order to reduce the burden that plastics pose to the environment. “Most plastics are fairly insensitive to moisture, which is part of the reasons for their long-life and persistence nature,” says VTT Technical Research Centre of Finland Research Professor Kristiina Kruus. Kruus will assume the deanship of the School of Chemical Engineering in December 2018. She heads the new Academy of Finland-funded flagship

project FinnCERES together with Aalto Professor Orlando Rojas. One goal of this multimillion-euro investment is to challenge plastics with new, plantderived materials, including those from lignocellulose. This is a major challenge, since plastics do have their strengths – and a considerable advantage for now.

Separate and build

The mass production of synthetic plastics began in the 1950s. Since then, a whopping 8.3 billion tonnes have been manufactured. Today, plastics are pro-

Finland has more forest biomass per capita than any other European country.

duced at a rate twenty times higher than in the 60s. Their popularity is easy to understand as plastics are waterproof, light and malleable, and are made from cheap raw materials derived from crude oil. “The amount invested in related infrastructure is so enormous that it is unrealistic to expect production to be wound down in the immediate future,” Kruus acknowledges. Side-stream hydrocarbons create the building blocks of plastics: long and strong synthetic polymer chains con-

sisting of thousands of molecules. For their part, lignocellulose biopolymers are created in the photosynthesis process through the interaction of sunlight, water and carbon dioxide, and they constitute the three main structural components of lignocellulose: cellulose, hemicellulose and lignin. “Our idea is to separate the constituent parts while preserving their polymeric structures, and then use these parts to engineer materials with entirely new properties,” Rojas says. This is already a familiar idea from AALTO UNIVERSITY MAGAZINE 23 \ 37

TRANSITIONING FROM THE PLASTICS AGE TO THE ERA OF BIOLOGICALS WOULD REVOLUTIONISE THE STRUCTURES OF INDUSTRY AND THE ECONOMY – IN A WAY THAT IS BENEFICIAL FOR FINLAND. papermaking, where wood fibres are first separated from one another and then mixed with water to form a slurry, which is then turned into paper after filtration, pressing and drying. Rojas and his colleagues want to go much further and, through novel separation and assembly strategies, to create functional biomaterials. Fibers, for example, can be deconstructed into nano- and micro-scaled objects, including nanocelluloses and nanolignins. In turn, they exhibit some typical features as those displayed by other nanomaterials, including a high reactivity and surface area. Such properties can be used in water purification, to remove harmful substances like pharmaceuticals and hormones or for air filtration. Assembly of colourless cellulosic nanomaterials in a specific way can result in coatings displaying vivid colours like those from the wings of butterflies, namely, structural colours. New functions, beyond colour, can emerge from such structures, which can be put to use. “We will go beyond the concept of ‘replacing plastics’. We already have materials that function just as well, we want to develop properties that are unique to biopolymers. FinnCERES offers a materials paradigm shift,” says Rojas.

“Finland has more forest biomass per capita than any other European country and we’re close to the top also on the global scale,” says Research Manager Katariina Torvinen of VTT. Finland exports some €14 billion worth of forest-based bioproducts each year. VTT calculates that the value of exports could up to double by 2050 by developing added-value products. This would boost GDP by around three percent. Strong growth like this would require Finland to become more than just a producer of wood pulp, but instead process biomass further into high valueadded products. The microplastics problem has already boosted the demand for cellulose fibres, and the search is on to develop alternatives to plastics also in the composite and packaging sectors. Torvinen predicts that the next item to be banned by the EU will be polystyrene, better known as Styrofoam. “In the USA, a few states have already banned its use. We’ve developed a technology that makes it possible to process biomass into a fibre-foam, which can replace Styrofoam. In addition, it can be used as heat and sound insulation in buildings, for example.” More than 300 million tonnes of plastics are produced annually in the world, while pulp production totals some 200 million tonnes. It is therefore not possiNature’s own composite ble to replace all plastics, nor is it the goal to do so, the researchers say. The smart The plastics industry of Europe alone choice is to utilise lignocellulose in prodgenerates more than €300 billion in annual turnover. Transitioning from ucts for which its natural properties are the plastics age to the era of biologicals best suited. would revolutionise the structures of Resilience and multi-functionality are industry and the economy – in a way some of the features in lignocelluloses. Lignocellulose is already a marvellous that is beneficial for Finland. There is no oil in Finland, but 78% of our land area composite material. “By evolution, the architecture of is covered by forest, and the amount of biomass keeps growing each year. lignocelluloses in trees allows them to 38 / AALTO UNIVERSITY MAGAZINE 23

FinnCERES – the biomaterials competence cluster • Academy of Finland-funded flagship project, which brings together the close research cooperation of Aalto University and VTT with industry as well as national and international research networks. • Based on scientific excellence, unique infrastructure and high impact. • Develops new bio-based materials for industrial-scale manufacturing for the international markets. These materials could be utilised both in everyday living and for industry: textiles, household goods, air and water purification systems, nextgeneration catalysts, ultralight transports and energy recovery systems. • The Academy of Finland has granted FinnCERES about €9 million in funding for the next four years. One-fifth of this will be spent on realising fresh openings suggested by the research community.

withstand storms, winds, flooding, heat, light and pathogens. From the strength point of view, wood can be processed to make bulletproof materials,” Orlando Rojas says. “Plastics are lifeless, made from fossil carbon, with no function on its own, unlike living materials that evolve to adapt to stresses of all types. These are the structures we want to learn from and to exploit, wisely and sustainably.” •


Equality as the theme of the art collection Percent for art principle was applied to Väre building. Text: Noora Stapleton Photo: Mikko Raskinen

Gloria Lauterbach hammered a large sheet of copper to form her sculpture on top of the concrete wall. THE CONSTRUCTION PROJECT, that

has characterised Otaniemi campus for over two years, was completed in summer 2018 upon the opening of Väre building. The new building gathers all departments of the School of Arts, Design and Architecture under one roof. Shopping centre A Bloc and the metro station are also in the same building block. Aalto University applied the percent for art principle in the construction project and acquired a vast public art collection for Väre. The art percent is a funding model for art purchases, where approximately one per cent of a building project’s funds are allocated to art purchases. The themes of the collection highlight Finnish equality, but also emphasize the strong international aspect of the school.

The works examine topics, such as, varying identities, nationality and refugee status as well as transgender issues and sexuality.

Artists with a strong connection to Aalto University The art collection consists of more than 25 pieces, and all of the artists have a connection to the university. There are pieces by alumni, former professors and students. The scale of materials used is vast: textiles, paintings, drawings and sculptures. The main works, chosen through an invitational competition, are placed in the cluster lobby areas of the building. Tommi Grönlund and Petteri Nisunen’s installation Insight com-

prises of rotating mirror surfaces that reflect what is going on in the surroundings. Kirsi Kaulanen’s sculpture Lumen is made of 150 laser-cut steel flowers and reflects the relationship humans have with nature. Gloria Lauterbach’s artwork Kreutz­ strasse won the student art competition, the goal of which was to generate ideas for decorating a 70-metre concrete wall next to Väre and the new main building of the School of Business. Lauterbach says that the artwork mimics the copper roof torn off by the storm Niklas in Zurich, Switzerland, in 2015.  “Over time, the copper roof will be covered with patina and a birch will grow from inside, representing silent, almost meditative change.” • AALTO UNIVERSITY MAGAZINE 23 \ 39


An unusually intelligent workmate Finnish company Curious AI is at the forefront in the race to develop artificial general intelligence. The company aims to create a self-learning digital workmate with the ability to work alongside a human expert. Text: Laura Siira Photo: Aleksi Poutanen IMAGINE BEING GIVEN an additional

ten personal assistants for your job. The assistants would be digital, and you could train them yourself for tasks of your choosing. What would you ask them to do? The artificial intelligence scene is currently engaged in intense competition on who will manage to develop the first so-called artificial general intelligence (AGI). Whereas existing AI technologies are developed for a specific need, such as facial recognition, for example, an AGI would, thanks to its human-like learning and deductive abilities, be capable of performing several tasks, with no programming skills needed to teach it. It would revolutionise how AI is applied in our everyday life. Established three years ago, Curious AI is one of the companies engaged in this race. It aims to create an autonomous and self-learning AI, which would be capable of working alongside an expert. It would, in other words, be a sort of digital workmate that anyone can train to be of assistance in various knowledge-intensive tasks. Last autumn, the company received almost three million euro in funding for its product development efforts. “An AI could, for example, learn to answer your e-mails or to manage your calendar notifications. Systems like this already exist, but someone is still required to code in all of their functions 40 / AALTO UNIVERSITY MAGAZINE 23

Senior Researcher Mathias Berglund envisions the future, but Curious AI’s offices in Helsinki are all retro.

by hand. The idea with AGI is that it’ll learn to use your calendar program without AI developers ever having laid eyes on it,” says Doctor of Science (Technology) Mathias Berglund, an Aalto graduate and one of the founders of Curious AI. Once given access to a company’s databases and internal exchanges of information, an AI could also propose alternative operating procedures for factories or bring together colleagues who are unknowingly working on the same subject. “The amount of knowledge work done around the world is enormous. If AI can boost its efficiency even by a sliver, the opportunities will be massive,” Berglund says. “And this is a conservative estimate of how AI could affect things.” Existing AI technologies are already capable of self-learning from data with the aid of deep neural networks. They will, however, only learn to do one specific task, and their learning is reflexive, which requires vast amounts of data. Doing things manually has, so far, formed a bottleneck in the development of AI. “A certain limit is reached in the case of self-driving cars, for example, because each situation needs to be taught beforehand. The advent of AGI would see cars take the next leap forward to become truly autonomous.” An AGI would not only be capable of learning from its environment, but also of planning its actions independently on the basis of this learning. Technologies for both already exist, but reconciling learning and the ability to plan has not been successful for now.

Users hold the levers of power

Users will rise to a more significant role in the future when anyone will be able to teach an AI. This also poses questions regarding the intentions and responsibilities of users. Up to now, AI developers have been experts like Berglund.

“The rise in user power can come with big risks attached when we consider, say, state actors – examples include killer robots used for military purposes or exploiting facial recognition and machine learning for the surveillance of citizens,” he says. And who will be responsible if an AI taught by you gives the wrong reply to an email addressed to you? Or if an AI causes damage in an industrial process? Or if a self-driving car causes an accident, whose fault is that? Berglund says this is a risk management issue: how freely will we allow AI to operate once its independent ability to reason increases. AI will inevitably make mistakes in the beginning. ”In principle, AI will do its best at what it is requested to do. However, we will certainly encounter situations that we, as the teachers of AI, have not thought to take into consideration, and that can have surprising consequences.” The deepest fear regarding AGI is that it might finally develop into a superintelligence that outsmarts humans, and break out of control by developing eversmarter technologies. AI singularity is the term used here. However, Berglund says the technology may not scale easily enough for this to be a realistic scenario. He compares development to forecasting the weather. ”Being able to add one more day to a two-day weather forecast does not mean that soon we’ll have the ability to forecast a whole month. One extra days requires an unbelievable amount of added computational capacity, and this slows down development,” Berglund says. “But it’s good that these things are being considered because the consequences could be catastrophic if realised.” Berglund thinks it is important to ensure that power and capital are not concentrated in too few hands in order to prevent the risks associated with AI from being realised. He points out that society

has an abundance of application targets for AI that can make the world, and our everyday lives, better. “Right now, commercial actors and universities alike are interested in solving largely the same questions associated with the development of AI.”

Towards human learning

Might the first company to develop an AGI be Curious AI from Finland? There’s no shortage of hype surrounding AI, but Berglund reckons there are only very few companies that can actually take the field forward. “A lot of components will have to fit together in order for AI to become really effective. International major corporations are developing technologies that are, as such, really useful, but still only good for a single purpose. Our advantages are a small enough team consisting of top international researchers as well as a long-term vision and plan for how things could function as a whole.” Berglund does not want to give an estimate of when every one of us can just pick a digital workmate off the shelf. The creation of an artificial general intelligence that learns like humans can be approached gradually. He is cautious when talking about more specific details. ”What we have already done to enable AI to learn from the world as well as learn to plan within that internal world and to make decisions based on this, is pretty revolutionary in fact. Development has come across certain challenges that we have solutions for. Others have not been able to do this.” And what would Berglund himself do if he had ten additional assistants in his work as an AI researcher? “I’d use them as my private tutors. They’d know what I am interested in and what level my current abilities are at, and could tailor the content of teaching to suit me precisely. I believe the importance of continuous learning will only increase in the future.” •



On science

Playing sparked an idea

how to transfer tacit knowledge Artificial intelligence could discreetly guide an employee to build and maintain equipment. Text: Riikka Hopiavaara Photo: Mikko Raskinen PROFESSOR YU XIAO and her 2 1/2-year-

old daughter sat on the floor and played with Lego. Her little fingers built a variety of structures which would have been difficult for someone else to replicate and make afterwards. Whilst watching her daughter play, Yu Xiao began to wonder if it would be possible to develop a solution that could learn from this kind of activity and automatically transfer the learning to others. Initially, the idea did not have an application, but after further research, Yu Xiao noticed that industrial companies had a problem to which her idea would fit. These days, instead of mass-produced items only, industrial equipment produce highly personalised products to meet the needs of different customers. The use and maintenance of equipment require expertise, which is often the result of

long-term practical work. However, teaching thousands of processes to new employees can be tricky. As Professor Xiao and her research team started to develop their smart and automatic tacit knowledge education system called ‘cognitive engine’, the Lego bricks transformed in her mind into a variety of industrial equipment and products.

Brain to brain learning

In practice, the system developed by the researchers involves a camera which records how an employee’s hands maintain or assemble a device. From the video the system learns the different work stages, how and in which order the various actions are taken, and the tools and components that are used. When an employee is working

on an unfamiliar device, he/she learns through various aids to work like an experienced employee. The technology monitors the employee’s actions and gives advice when he/ she is about to make an error. Time and effort is saved when there is no need for an instructor to be present when teaching new skills. Reliability is also increased when errors cannot be made without the system noticing them. The employee’s stress levels are also reduce as they do not need to worry about making mistakes. “Practical implementation is still being planned, but we intend to use a number of techniques, such as wearable glasses, smart gloves, and various sensors”, says Yu Xiao. The purpose is to create a learning experience in which the employee is guided without having to follow written instructions. “The guidance is invisible, so it does not irritate.” The project utilises expertise in many different fields, from basic research to artificial intelligence, smart clothing and user experience. The same concept can be utilised in all areas requiring cognitive abilities, even in the teaching of bus drivers or pilots. Professor Yu Xiao applied for and was granted funding, and now, one year and a half after an idea borne out of playing with Lego, the prototype is in the pipeline. She is confident about the future. “I believe that in about a year or so we will have a complete prototype and the product can be commercialised.” •

Cognitive Engine for Assembly and Maintenance Automation, also known as the CEAMA project, received €700 000 in Research for Business funding.



Student wins major award with mobile music app “Openness is the way forward,” says Ilari Niitamo, winner of the Apple Design Award 2018. Text and portrait photo: Lucas Millheim Photo: Yatatoy

“I AM AN MA STUDENT in New Media,

set to graduate in December 2018. Our company, Yatatoy, is a bit unique. It’s more of a creative collective than a company company. Lucas Zanotto, the founder of Yatatoy and myself were both working on another project for Hello Ruby, which I also just wrote my MA thesis about. We worked on Bandimal for all of 2017 and released it in November. It’s a music composer for kids. You can set up a drum loop, swipe through animals to change instruments, compose melodies, change loop speeds and add effects. Kids like the animal animations that change as they make music. We tried to make the app super easy and fun for kids to use. Our audio wizard, Ulrich Troyer, would design, compose and produce the music system and samples in Vienna. Lucas and I both worked in Helsinki. Lucas created, illustrated and animated the animals while collaborating with Ulrich to get the stylization of the instruments correct. Lucas and I would collaborate on the app design and user experience, and I would build the app. Bandimal was programmed using Apple’s native technologies. Now we got this prize! The Apple Design Awards is an annual award ceremony, held at their yearly WWDC developer conference that celebrates great-quality apps. Essentially, they give an award to the ten best apps of the year, in their view. It’s arguably the biggest and most meaningful award you can get in the mobile business. The awards ceremony was intense. On the first day they gave me a guest pass,

The Bandimal application is a music composer for children.

and I was seated in the front during the keynote talks. When the ceremony part started, we got called up onstage as the first app to be awarded. The rest of the conference was incredible: I had interviews with media, I got to meet Apple executives at a special champagne reception, but people also came up and congratulated me out of the blue. Afterwards, it was the most popular app in Finland, and number 1 in the Kids category in around 60 countries and top 10 of all apps in 25 countries. The success of Finland in app or game development has something to do with our background in technology and mobile phones, I believe. There is a basic understanding about what people need in the mobile space, what is practical and pragmatic. Also, I believe Nordic minimalism in design plays a role in this success, especially the light aesthetic touch that is common.

In the Finnish psyche there is a certain quirkiness and underdog mentality, and I think that also adds a bit of creativity to what we do. Also, we are a small country and circles here are small; you could almost say everyone knows everybody else. It’s easy to connect with people and get things done. I believe the way forward is to embrace openness in a completely new way. The tech scene in San Francisco has done that already, they’ve gotten past wanting to keep things to themselves. The old way of thinking was: be careful how much training you give your employee, they might learn too much and leave you. Don’t talk to competitors because you might give away secrets. When I was working in San Francisco and we came across a technology that was especially useful, we would just call the CEO of the company and go out to lunch with him. Then we would ask directly for help integrating it into our own system. I think openness is the way forward.”• AALTO UNIVERSITY MAGAZINE 23 \ 43

Current affairs

Learning to learn and other trends Aalto University will begin preparing a new strategy next year. To form a basis for this work, students were asked to explore what kinds of megatrends would affect the university’s future. The review extended as far into the future as 2035. Text: Laura Siira Photo: Anna Berg MOST REVIEWS of the future predict that

the pace at which the world is changing will increase. Unprecedented evolutionary forces are, in the wake of technological development, expected to target especially working life and the field of education that prepares people for it. What kind of a future is Aalto University getting ready for in its upcoming strategy? What trends will affect teaching and research, and, on the other hand, the existing and future Aalto community? Students were asked to pitch in the strategy preparation process, which is set to commence fully in 2019. Based on a review of various publications and expert interviews, a student group steered by Management Studies Professor Nina Granqvist defined five central trends that will affect the future of Aalto. “Megatrends will bring changes to the university’s operating practices either directly or indirectly. Automation, for example, will increase the need to retrain the workforce, which in turn will boost demand for more flexible studies and add to competition in the educational sector,” says School of Business student Joona Orpana.

Soft skills emphasised

The need for lifelong learning is a trend, which is in many ways changing not only the educational sector, but also our everyday lives. Debate often focuses on the mastering of technical skills, but the interviews conducted by the working group also raised the growing significance of so-called soft skills like creativity. “Technology is developing constantly, and recently acquired technical skills A group of students representing the different fields of Aalto delved into the world of megatrends. The participants were Jacqueline German, Joona Orpana, Viktorija Piaulokaite and Andreas Gratz. 44 / AALTO UNIVERSITY MAGAZINE 23

can only underwrite your competence for a little while. This is why the ability to learn new things, to understand yourself and others as well as to strive for change that accords with your own values are emphasised. Sustainable competence is created by combining hard and soft skills in learning,” says Orpana. The students and Granqvist all point out that the future is hard to predict. Megatrends are abstract, complex totalities, and it is challenging to foresee what combined effects they may have. “It is important for the strategy work to consider the university’s vision as well as the plans employed to achieve it – but these plans must be flexible at all times. Evaluating the future is a continuous effort you engage in alongside day-to-day development work,” Granqvist says.

Involving the community

Future strategy workshops and community discussions will benefit from the work done by the students. Community involvement is a central aspect of Aalto’s strategy work, says Development Director Sirkku Linna. “We want to make the process as open and transparent as possible – something that’s easy to get onboard with,” says Linna. “We are not just preparing to adapt to a changing world, we’re getting ready to shape the future into what we want it to be. Aalto is in a key position to, for example, develop technological breakthroughs and steer how technology shapes our reality.”•

Five megatrends affecting the operating environment of Aalto

Disruptive technology mindset

The current pace of technological change is unprecedented. The exploitation of big data, artificial intelligence and quantum computers are shaping reality thoroughly. One way or another, life will revolve around technological upheavals. “Members of the future Aalto community could avail of personal AIs in their studies, work and daily life. It might be possible to provide, for example, tailored teaching with their aid.”

Radical business-university interdependence

The field of education is going through a transition. The need for continuous education is growing, while digitisation is also opening fresh opportunities for evolving markets. New actors, which will challenge the traditional role of universities, will enter the field of education. Universities must establish even stronger relationships with business in order to ensure that their training remains current as well as to safeguard their funding. Education and research will more often be arranged in cooperation between different actors. “Resources could be divided more diversely in the ecosystems that will grow between universities and businesses. Companies could have access to Aalto courses and research data, and, in return, corporate personnel would donate their time to the University by mentoring and giving lectures.”

The quotes are pulled from the review written by the students.

Imperative environmental action

Citizens, companies and states must all demonstrate greater responsibility for curbing climate change. Fundamental societal changes, which can shape the power structures that steer the global economy and encourage environmentally sound innovation, are needed. “Aalto has the opportunity to become worldfamous as a promoter of sustainable solutions and wellbeing. Research activity is already focusing on many fields of central significance to the environment, like the circular economy as well as alternative energy and materials solutions.”

Changing power configurations

The impact multinational corporations have on political decisions will increase, while, on the other hand, people’s interest in traditional participation will decrease. Inequality will grow at the same time. Democratic institutions will find themselves in a challenging situation. New ways to wield influence also on the global level will be needed. “Global challenges could be tackled on new types of political arenas, where decision-making is not dictated by states. Aalto’s entrepreneurship ecosystem, for example, could serve as one.”

Long life purpose

The changes brought about by technological development and, in turn, lengthening lifespans translate into a continuous need to learn new things and update your knowledge. People’s roles and work tasks will change more frequently, and here education can play a part in providing meaning and significance to life. “Demand for, for example, multidisciplinary courses, expertise in soft skills and flexible studies that adjust according to work will grow. Alumni will return to the sphere of the university several times during their lives, which is a great opportunity to also utilise their competence in mentoring, for example.”


Doctoral theses

A revolutionary quantum state Otto Mustonen was the first scientist in the world to succeed in producing a quantum state predicted by a Nobel laureate in physics. This is a promising step toward super-efficient quantum computers.

measurements under the direction of Braunschweig University of Technology Professor F. Jochen Litterst. The measurements utilised muon spin spectroscopy, in which very short-lived elementary particles – muons – interact with the material being researched. The quantum spin liquid formed in a cooled material. Detecting the phenomenon required substantial items of equipment including a large particle collider. The hours of each day spent on measureText: Marjukka Puolakka ment had to be utilised very carefully as the daily cost of using such equipment is Photos: Jaakko Kahilaniemi in the tens of thousands of euro. THE OPPONENT, University of Oxford “We worked 16-hour days for four days Professor Peter Battle, considered Otto in a row. We measured several samples, Mustonen’s thesis one of the best docand there were some surprises. It was toral dissertations he had seen during his a superb feeling to then see the data dementire career. But the path to the public onstrate that we’d succeeded in creating examination event in which his work cul- quantum spin liquid.” minated included many twists and turns. The production of quantum spin liquid “Chemical intuition formed the backwas made possible by a method of tailorground to the study – what happens when ing the magnetism of materials that was you bring certain elements together? But developed at Aalto University. in research you can’t plan what will and “The method can be utilised with what won’t work for four years ahead,” many other materials as well, and could says Mustonen, who is grateful for Aalto in future result in something really interUniversity’s funding, which allowed esting.” certain freedoms for defining the area of research. Scotch and bad movies In 1987, physicist and Nobel Prize Following the completion of his doctorlaureate Philip W. Anderson proposed ate, Mustonen will be researching new that high-temperature superconductivmaterials and magnetic states at the Uniity, i.e. the loss of electrical resistance, is related to a specific quantum state called quantum spin liquid. In this state, the electrons of a material behave like a liquid and will not order even at absolute zero. Quantum spin liquids are promising materials for so-called topological quantum computers, which could perform extremely fast computations. The success of the doctoral research required the close cooperation of several chemists and physicists. In addition to Aalto, research groups from the UK, Germany, France and Brazil took part in the project. “I’m grateful to Professor Maarit Karppinen for this cooperation, which allowed me to establish contacts with other universities and gain access to the measuring devices of the best groups in the field.”

Sixteen-hour days

Mustonen experienced his most magnificent moments at the Paul Scherrer Institute in Switzerland, when he performed 46 / AALTO UNIVERSITY MAGAZINE 23

versity of Strathclyde, Glasgow, to which he has been awarded a three-year grant. In Scotland, he expects to also enjoy the amazing nature, bewildering Scottish accent and fine whiskies, a hobby of his inspired by a Welsh colleague during post-graduate studies. “I’ve already located the nearest distilleries, as many workmates from Aalto are planning to visit me.” In addition to whisky, the team at his former workplace was united by a keen interest in bad movies, especially lowbudget action flicks from the late 1980s that were so ubiquitous during the golden age of video rental shops. “It is so much fun to watch and comment on a film together when its production is beyond poor and the events and dialogue make absolutely no sense.” The work community helped him get through the doctoral project in many ways, and Mustonen has faith that his new workmates will also include fine characters. Quality films you can enjoy by yourself, but you need a good crowd for a real turkey. Otto Mustonen 25.5.2018: Strongly Correlated Oxides: Half-metallicity in Chromium-based Rutiles and Quantum Magnetism in Copper-based Double Perovskites.

Doctoral theses

All doctoral theses online: aaltodoc.aalto.fi, shop.aalto.fi

Large energy companies in transition LARGE ENERGY companies face ever increasing

Microlitter a challenge to wastewater treatment plants OVER THE PAST few decades, there has been more and

more concern about microlitter and especially microplastics (sized 0.02–5 mm) polluting the environment. In her doctoral thesis, M. Sc. Julia Talvitie examined the role of wastewater treatment plants (WWTPs) as sources and pathways of microlitter entering the aquatic environment. This is the first doctoral thesis in Finland to focus on microplastics. A conventional Finnish WWTP can efficiently remove (> 99%) of the microlitter sized 0.02–5 mm that arrives at the plant. As vast volumes of wastewater are constantly discharged into the Baltic Sea and Finnish lakes and rivers, the role of WWTPs as microlitter pathways may nevertheless be significant. Finnish WWTPs annually discharge an estimated 480 billion microplastic particles into natural waters. Baltic blue mussels collected from a wastewater receiving area had higher microlitter concentrations than those from a reference site. The removal of microplastics can be further enhanced by advanced final-stage wastewater treatment technologies. The membrane bioreactor Talvitie studied removed an additional 99.9% of microplastics during treatment. Furthermore, sand filtration and other effluent tertiary treatments were found to efficiently remove microlitter. The results of the research are useful in designing future wastewater treatments plants for optimal overall waste water purification efficiency. This will result in less microlitter entering the aquatic environment.

pressures to tackle the challenges of climate change and technological development. The doctoral thesis of M.Sc. Eeva-Lotta Apajalahti studied the activities and roles played by Finnish energy companies in the transition of the energy system towards renewable energy. A large municipal energy company has, during its century-long existence, developed a deep carbon and technological lock-in. Underlying this is the company’s large-scale combined heat and power (CHP) production. To break free of such a lock-in is challenging, yet not impossible. The company has already decided to close down one of its coal power plants before the end of its estimated useful life. Three pressure factors are behind this: the need to reduce coal use, CHP production’s profitability challenges and the desire to free up city space for residential use. The transition of our energy system is not easy because of its rigidity, which is caused by past commitments to technologies, infrastructures, organizational forms and resources, Apajalahti finds. However, by entering the renewable energy business, large energy companies may push the energy transition into a totally new level – they have the resources to reshape, for example, the technological fields of solar power and electric vehicles. The companies may become bridge builders of transition that creatively integrate new and existing technologies and infrastructure. • Eeva-Lotta Apajalahti 8.6.2018: Large energy companies in transition – from gatekeepers to bridge builders.

Julia Talvitie 18.5.2018: Wastewater treatment plants as pathways of microlitter to the aquatic environment. AALTO UNIVERSITY MAGAZINE 23 \ 47

Everyday choices

Petri Myllymäki, is AI a threat to humanity?

The professor fired his smart helper and loses no sleep over super AI. Text: Paula Haikarainen Photo: Jan Ahlstedt You’re in charge of the Helsinki Institute for Information Technology (HIIT), a joint Aalto University and University of Helsinki research institution. Do you spend more of your working day in Helsinki or in Espoo? I’ll quote the famed probability theorist, ski jumper Matti Nykänen, who said the odds are fifty-sixty. Even though it sometimes feels like I spend most of my day on the metro, meetings with cooperation partners and funders usually take place in Helsinki.

An assistant could filter the most urgent and important email messages to the top, and arrange even those large meetings by finding suitable times and places. This would save enormous amounts of time. Your work takes you around the world to conferences and research meetings. You must accumulate a lot of travel days. I don’t spend as much time travelling nowadays as I did back when I worked as a full-time researcher, but I do try and make it to a top conference at least once a year just to keep up with developments in the field. The Uncertainty in Artificial Intelligence conference is perhaps the closest to my own field of research, and a sort of “home conference” which I’ve attended 28 times in a row. I received an award on my 25th year, so I probably hold some kind of world record now.

There are now dozens of professors focusing on AI research in the Helsinki region, the volume is substantial even in international comparison. The new FCAI competence center will see us step proudly into the limelight to ask who wants to join us in cooperation.

Super AI refers to machines that are more intelligent than humans. This sounds scary, but researchers disagree on whether or not it is possible to build such a thing. What do you think? If it is possible Did you always want to become to create it, I’d rate the likely threat on a professor? No. I dreamed about being a similar level to the possibility of CERN a researcher, but became a professor blowing up the planet with an antimatter instead – a jack of all trades and particle. Or that all life would be extinan administrator. guished by nuclear war. It’s not worth losing sleep over. How is AI a part of your everyday I’m put off by clickbait news declarlife? Every one of us uses AI. For examing that AI has beat the world’s best Go ple, when you access your email, AI has player or the world champion of chess filtered the spam, preferably with suc– and now it can drive a car, too. We cess. Online search engines are also Is distance participation not should keep in mind that the AI that’s examples of primitive AI. an option? I think the most important best at Go doesn’t even know that cars or I have, of course, experimented with part of a conference is meeting people humans exist. And it isn’t thinking about using a speech-recognising AI assistant. and establishing contacts. Many a project destroying humanity in its free time. The limits of its intelligence were swiftly has gotten started from a discussion had We are certainly safe from that. encountered. When I tried to arrange over lunch or during a coffee break. For now, AIs are built to do or solve a meeting in Helsinki with a person a single thing. Self-driving cars are an who lives in America, the assistant Which part of the world is furthest in example of this. If you were to fall down managed to schedule lunch in both of development? Asia is steaming ahead on a road, a robot car would swerve our calendars, but it used San Francisco right now, and it’ll be exciting to see if the around you because it spots an obstacle time. It did manage to change the time lead passes from North America to Asia. on the road. But it wouldn’t have after I remarked a couple of times that There hasn’t been much noise about the understanding to call an ambulance. the meeting is set for Helsinki, but this, but Finland, especially the Helsinki It could, of course, be taught to raise I lost interest in the tool following region, is an internationally significant the alarm if it sees a person lying on the experience. centre of AI. We’re held in high regard the ground. But it is hard to list every Every day I think about what – that’s evident at any event that hosts single possible variable – and it’s not a machine could do on my behalf. peer-reviewed and invitational speeches really possible to teach common sense. When I cycle my electric bike to work, on this field. You don’t see as many peoEveryday things that we take for granted I wonder why it doesn’t yet have a brake ple from, for example, the other Nordic can be insurmountably difficult for assist system to help in emergencies. countries at them. a machine. • Or, when I take public transport, I think This is thanks to certain powerful the journey planner could be more intel- personalities like Teuvo Kohonen, ligent. At present, it can monitor the situ- Erkki Oja and Heikki Mannila. Fresh The Fat Lizard restaurant in Otaniemi ation and schedules in real time, but does science has sprouted from the foundais an important meeting spot for Petri Myllymäki. “I host meetings, not estimate the uncertainties of various tion built by such pioneers, and it keeps routes or calculate probabilities. getting stronger. even development discussions, here.” 48 / AALTO UNIVERSITY MAGAZINE 23


More than a backdrop to the music From short films to arcade video games, Flow Festival was awash with creative work by Aalto students.

VIDEO: AaltoxFlow http://bit.ly/2P4LJ3D

Text: Lucas Millheim Photos on the inside covers: Anni Kääriä, Mikko Raskinen

Thu 1 Nov 2018, 11 - 17 Otahalli, Otaranta 6, Espoo


the once-bleak industrial space between Helsinki’s Hanasaari power plant and the Kalasatama construction site is transformed by design, art and creativity for over eighty thousand music lovers. Among the attractions are works by over 40 Aalto University students. The most visible work is the Aika-lava stage, which debuted at last year’s Pori SuomiAreena, and was now set up on the festival site. Skateboarders immediately repurposed the building even before the festival. “It’s nice to see it here in these sur-

roundings and to see people using it for something new,” says Ransu Helenius, producer at the Aalto University Wood Program. Inside the wooden structure, over 100 square metres of flowing white fabric was hung to create a dreamlike environment where concertgoers could retreat from the festival bustle – and play arcade games built by students. “We made games that related to the theme of the Flow Festival, which is sustainability and the Baltic Sea,” says Jung Huh, an exchange student from Korea.

As for Caleb Rugg, who built the Global Warning arcade-style game for the festival, it’s all about the exposure: “Mostly, I’m looking forward to having thousands of people play our game.” Students also painted a 100-metre mural wall and created animations for the main stage screens. Short student films were shown to packed audiences in a separate screening area. On Sunday, students organised a communal art workshop with recycled fabrics, while the Aalto Junior crew made giant soap bubbles and painted canvas bags with kids. •

Thu 1 Nov 2018, 11 - 17 Otahalli, Otaranta 6, Espoo Career Forum area argest ICT

Recruitment ndLargest CareerICT Fair Recruitment n Finland with Career Fair verand 100 exhibitors in Finland with


over 100 exhibitors Thu 2018,11 11--17 17 Thu11Nov Nov 2018, Otahalli, Otaranta Otaranta 6, Otahalli, 6,Espoo Espoo

• Interview Coaching • CV & LinkedIn Clinic Career Forum area CareerResume Forum area Video ••Interview Coaching • Interview Coaching ••CV & LinkedIn Clinic Photo Station • CV & LinkedIn Clinic ••Video Resume Snap Corner • VideoMeeting Resume • Photo Station • Photo Station • Snap Meeting Corner

New TalentX area • Snap Meeting Corner

New TalentX area • Test your skills by solvin New TalentX area

• Test your skills solving life practical andby • Test your skills bywork solving entrance practical and work work life related problems and life Largest ICT RecruitmentCareer and Career Fair practical Forum area Largest ICT Free entrance related problems • Interview Coaching related problems inRecruitment Finland with over 100 exhibitors

and Career Fair in Finland with over 100 exhibitors

• • • •

https://fairs.aalto.fi/talentit/ https://fairs.aalto.fi/talentit/

CV & LinkedIn Clinic Video Resume Photo Station Snap Meeting Corner

@TalentITFair New TalentX area @TalentITFair • Test your skills by solving practical and work life related problems

Free entrance Organized by Aalto University, Organized by Aalto University,

Advancement andUniversity, Corporate Engagement Organized by Aalto Advancement and Corporate Engagement Advancement and Corporate Engagement https://fairs.aalto.fi/talentit/

@TalentITFair @TalentITFair



TalentX Area in TalentX area in TalentX in cooperation with Area AYY cooperation with AYY

cooperation with AYY


artificial intelligence atomic layer deposition technology biomaterials


Profile for Aalto University

Aalto University Magazine 23 – English edition  

This issue is about artificial intelligence.

Aalto University Magazine 23 – English edition  

This issue is about artificial intelligence.