Can computers create new songs?

Can computers create novel songs? At what point can computer software be called creative?

A team or researchers led by University of Helsinki and FCAI professor Hannu Toivonen, tackle these questions in their newly-published paper in Connection Science. They argue that a crucial element of creativity in software is its ability to self-monitor and self-modify its own operation. This ability is known as transformational creativity.

The research address many core topics of artificial intelligence: self-awareness, self-adaptation, and creativity of intelligent software.

"Our work furthers the explainability of AI and the ways intelligent systems and users can interact," says professor Toivonen.

In the paper, Toivonen and his colleagues provide a concrete and implemented architecture for transformation creation of songs. 

See the article: Jukka M. Toivanen, Matti Järvisalo, Olli Alm, Dan Ventura, Martti Vainio & Hannu Toivonen (2018), Towards transformational creation of novel songs, Connection Science, DOI: 10.1080/09540091.2018.1443320.


AI application for treatment of gestational diabetes

AI allows individualized predictions for expectant mothers and newborn children. The aim of the individual recommendations is a positive experience for the user combined with activity that is beneficial for the glucose level.

About 52,000 women give birth in Finland every year, and 18 per cent of them – nearly 10,000 – are diagnosed with gestational diabetes. Of these, roughly half develop type 2 diabetes later on.

CleverHealth Network, an ecosystem coordinated by the Hospital District of Helsinki and Uusimaa (HUS), is now launching its first development project with funding granted by Business Finland. The main partners in the gestational diabetes project are HUS, Aalto University, the University of Helsinki, Elisa and Fujitsu. The project is run by FCAI researchers Pekka Marttinen ja Giulio Jacucci.

The project aims to improve the treatment and monitoring of gestational diabetes by developing a mobile application for measuring the mother’s blood glucose levels, physical activity, nutrition, pulse and daily weight and storing it in the cloud in real time.

Rid of routine coding – AI automates the construction of large information systems

A new technology under development in Aalto University will bring down development costs so much that complex systems can be built from scratch.

Business Finland has granted 678,000 euros to a team lead by Aalto University’s Jussi Rintanen for the commercialisation of a new information system technology based on artificial intelligence. Rintanen wants not only to automate the development of large information systems but also to integrate all parts of software development into a single functioning whole.

The Finnish government: AI will be the new normal

Last year, the Finnish Ministry of Economic Affairs set up a working group to find out how Finland can and should commit to be in the forefront of artificial intelligence in industry, business and society at large.

The report and more information of the group's work is now available in English:
Download the entire report here as pdf.

In sum, the report presents eight key actions that guide Finland to the AI age.

Enhancement of business competitiveness through the use of AI
Effective utilisation of data in all sectors
Ensure AI can be adopted more quickly and easily
Ensure top-level expertise and attract top experts
Make bold decisions and investments
Build the world’s best public services
Establish new models for collaboration
Make Finland a frontrunner in AI

Aalto University #19 in the world and #4 in Europe in Human–Computer Interaction research

In the five-year period 2013–2018, Aalto University ranks #19 globally and #4 in Europe in Human–Computer Interaction research. In the past ten years (2008–18), Aalto’s the respective rankings are #22 # and #4.

The rankings are made by that pulls its data from the DBLP Computer Science Bibliography. The rankings are based on the volume of publications in each institution.

See full rankings here. 

FCAI partner universities are recruiting in AI – see open positions

State-of-the-art research and knowledge on AI and closely related fields are in great demand in Finland and in FCAI partner universities. Currently, there are several calls for everyone from exceptional doctoral students and postdocs to professors. See listing below for individual calls.

University of Helsinki
Four (4) professor positions in Data Science

Aalto University                                                                                                                        Professor in Artificial Intelligence (tenure track)
Review begins 25 January 2018; position open until filled.

Professor in Machine learning (tenure track or tenured)
NOTE: call not open anymore.

Aalto University and University of Helsinki
Several Postdoctoral Researcher and Research Fellow Positions in ICT

Linking two or more of the following themes:

  • artificial intelligence and machine learning
  • data science
  • privacy and security
  • computational health
  • human-computer interaction

Helsinki ICT network (Aalto University and University of Helsinki)
Positions for Exceptional Doctoral Students (NOTE: deadline 31 Jan 2018)

Helsinki Institute for Information Technology HIIT & Institute of Molecular Medicine Finland FIMM
Five (5) open doctoral student positions



StanCon 2018 will be at Aalto University

The StanCon 2018 Conference on statistical modeling and probabilistic
programming will be in Helsinki at Aalto University 29–31 August.

Stan is a statistical modeling language used by thousands of scientists, 
engineers, and other researchers for statistical modeling, data analysis, and
prediction. It is being applied academically and commercially across fields as
diverse as ecology, pharmacometrics, physics, political science, finance and
econometrics, professional sports, real estate, publishing, recommender
systems, and educational testing.

See more info about, among other things, talks and poster submissions at:

Artificial photos created by two neural networks training together

A research group at NVIDIA trained neural network algorithms to teach each other to create artificial but photorealistic images of human faces. The team features FCAI professor Jaakko Lehtinen.

Read more here (New York Times).

Or if hit by the NYT paywall, here's Wired and professor Lehtinen on the subject:

Also a story in The Verge:

Artificial Intelligence Day attracted 600 people fascinated by AI

The Artificial Intelligence Day brought together researchers, companies and the public sector involved in the fast-developing field of AI. “Without great science, there cannot be any innovations in industry,” summed up one company participant, Kimmo Pentikäinen, Elisa’s vice president of business development.

On Articial Intelligence Day on 13 December 2017, 600 artificial intelligence experts and enthusiasts gathered in Dipoli, Aalto University’s newly-renovated main building.

The organiser, the new Finnish Center for Artificial Intelligence FCAI established by Aalto University and the University of Helsinki, wished to promote matchmaking, information sharing and cross-border collaboration with the event.

Unique research problems add value to both academia and companies

Representatives of over 180 companies were offered matchmaking opportunities during pitching, demo and poster sessions. One of the large Finnish companies present was Elisa, who’s vice president in business development Kimmo Pentikäinen met up with Samuel Kaski, Professor at Aalto University and Head of FCAI, in the AI Day networking area

They discussed the needs of Elisa as an eager partner for research institutions.

“It’s a massive amount of mobile data that we have available in networks, and what we are always aiming for is to provide this data for scientific purposes,” Pentikäinen said. “That’s the start of the research collaboration: for us to know what we do not know yet. And that’s the essence of doing scientific research, right?”

Samuel Kaski highlighted the importance of finding unique research problems that would add value to both the research community and companies.

“What’s most fruitful for both is to identify things that no one else is looking for yet, and try to find a match for our methods and actually solve problems that you need to get solved but haven’t realised that yet. Maybe that could be the kind of killer match we want to find!” Samuel Kaski envisioned.

“I couldn’t agree more because what we are seeking from academic research is the things we do not know,” Kimmo Pentikäinen concurred. “That’s how we can identify the next big things that will change our own operations, the entire industry, and society as a whole in a fundamental way.”

Strong interest for applications and cross-field collaboration

In the conference feedback that FCAI collected, several AI Day participants expressed their interest for hearing more about concrete AI applications and business opportunities. They would provide deeper insight for technology experts, company executives and decision makers.

In addition, more information on cross-border collaboration between technology, humanities and creative fields should be visible and disseminated in the future AI Day events in order for AI to address wider societal problems.

Participants also listed topics that would interest them in the forthcoming AI events: applications of AI in industrial settings, traffic, and data management, AI and data security, interactive technologies, and women as AI developers.

FCAI is planning to make AI Day an annual event. The next seminar is scheduled to be held in the autumn of 2018.

Further information:
Terhi Kajaste, FCAI Corporate Liaison, Aalto University

Photos: Matti Ahlgren / Aalto University

Six papers by FCAI researchers at NIPS 2017

At the Conference and Workshop on Neural Information Processing Systems NIPS 2017 in California in December, FCAI researchers presented altogether six papers. The 2017 conference broke all previous attendance records which in itself is a clear sign of the booming and wide-spread interest on artificial intelligence research.

The presented papers at NIPS 2017:

Kari Rantanen, Antti Hyttinen, Matti Järvisalo
Learning Chordal Markov Networks via Branch and Bound

Sami Remes, Markus Heinonen, Samuel Kaski
Non-Stationary Spectral Kernels

Mikko Heikkilä, Eemil Lagerspetz, Samuel Kaski, Kana Shimizu, Sasu Tarkoma, Antti Honkela
Differentially private Bayesian learning on distributed data

Isabeau Prémont-Schwarz, Alexander Ilin, Tele Hao, Antti Rasmus, Rinu Boney, Harri Valpola
Recurrent Ladder Networks

Antti Tarvainen, Harri Valpola
Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

Kiran Garimella, Aristides Gionis, Nikos Parotsidis, Nikolaj Tatti
Balancing information exposure in social networks