A sample of the activities of the HIIT programmes is presented here as the research highlights of 2018. In addition, HIIT supported initiatives of the broader Helsinki ICT Community, as described in Community Support.
Teemu Roos emphasises the role of universities in realising the benefits of AI
Artificial Intelligence is expected to provide solutions to a wide variety of problems and needs, but in order to realise the benefits while minimising the damage, long-term research and broad access to education are needed, says Teemu Roos, associate professor at the Department of Computer Science at the University of Helsinki.
Much of artificial intelligence research is multidisciplinary. It is often slow and requires long-term funding.
“It may take years for people from various fields to learn to talk with each other. Projects often last a couple of years, and launching them is risky, if there is no certainty about continuing funding,” says Roos.
He thinks that universities receive too little recognition for the artificial intelligence research they conduct.
“There is a great deal of talk about research conducted by Google, IBM and Facebook. Yet the individuals working in these companies have been educated by universities. Secondly, without the university ecosystem, the majority of companies could not utilise artificial intelligence. Large companies may conduct their own research and product development, but even they don’t have the desire or resources to conduct the critical basic research on which innovations are based.”
AI being a hot topic, decision-makers in politics and the corporate world easily lose sight of the difference between experts and “experts”. Genuine expertise is needed to ensure that the Finnish population can be trained to recognise the possibilities and dangers inherent in artificial intelligence – for example, the potential to shape opinions.
“When it comes to issues related to artificial intelligence, researchers are the experts you should listen to,” says Roos.
These thoughts are well in line with FCAI’s mission to create Real AI for Real People in the Real World. In addition to his distinguished research in machine learning, Roos himself is the lead instructor of the Elements of AI online course that aims to educate 1% of the Finnish population to understand the basics of artificial intelligence.
For the full interview, please see the University of Helsinki website.
Finnish machine learning research behind acclaimed acquisition deal
Ekahau, the leading solution provider for enterprise wireless network design and troubleshooting, has been acquired by Ookla. Artturi Tarjanne, a general partner of Nexit Ventures, thanks in his blog the “super talented CoSCo research team behind innovations and magic of Ekahau technology” (see Nexit’s Best Exit Ever). The researchers behind the technology that led to the establishment of Ekahau in 2000 were Henry Tirri, Petri Myllymäki, Teemu Roos, Kimmo Valtonen, Tomi Silander, Petri Kontkanen, Antti Tuominen, Jussi Lahtinen and Hannes Wettig, forming the core of the CoSCo research group of the University of Helsinki and HIIT at the time.
The size of the deal is not public information, but Helsingin Sanomat, the main Finnish newspaper, estimates it to be in the range of 100-130 M€. Petri Myllymäki, who is currently the Director of HIIT and vice-director of the Finnish Center for Artificial Intelligence, comments: “After the 18 long years when Ekahau had to first face the burst of the IT bubble in 2001, and then later struggle with problems in finding the right business model, I was very happy to hear these good news about the company that started on ideas based on our long-term basic research in machine learning. As the story in Helsingin Sanomat says, the journey was unfortunately so long and winding that for the original innovators this is no longer a great personal financial success, but this is still a great a success story for Finnish science and more widely for the Finnish society: according to Tarjanne, the majority of the 75 professionals employed by Ekahau are located in Finland, so the impact on local high-tech employment alone has already been substantial. This is a prime example of what top-level Finnish AI research can lead to. Kudos to the Nexit and Ekahau teams for not losing their faith and for all the hard work they have done during these years!”
AI Day 2018
On 12 December, AI Day 2018 took place at Aalto University Otaniemi campus. The event featured talks, panel discussions and networking events with leading researchers, companies and policy makers across the field of artificial intelligence and machine learning.
Over 500 people registered for the day, which was attended by over 20 companies and public organization, including Supercell, Nokia and OP Bank. The event had representatives from around 150 organizations present.
“This looks very promising for Finland – this time we had much more research content in the talks and the number of people from companies interested is still strong!” said FCAI director, Professor Samuel Kaski. He added “We had 543 registered participants, roughly the same number as last year, even though now we had a small entrance fee to cover the costs. So people seem not to be coming just for the free coffee.”
As well as delicious coffee, visitors the event were treated to opening talks by Ilona Lundström, the Director General of the Ministry of Economic Affairs and Employment. This was followed by parallel sessions on Understandability in AI chaired by Professor Antti Oulasvirta from Aalto, and Privacy, Security and Fairness in AI chaired by Professor Antti Honkela from University of Helsinki.
The afternoon featured a panel Chaired by Risto Nieminen, the President of the Finnish Academy of Science and Letters, made up of Meeri Haataja, CEO and cofounder of Saidot.ai; Professor Petri Myllymäki, director of Helsinki Instituted for Information Technology, Professor Aki Vehtari from Aalto University, and Professor Petri Ylikoski from university of Helsinki, who were discussing the Scoeital Impact of AI. Running in parallel was a series of academic talks on Data efficiency in AI, chaired by Professor Alexander Illin from Aalto.
Building an AI to predict if you carry a killer on your skin
Staphylococcus epidermidis is an ubiquitous colonizer of healthy human skin, but it is also a notorious source of serious nosocomial infections with indwelling devices and surgical procedures such as hip replacements. It has not been known whether all members of the S. epidermidis population colonizing the skin asymptomatically are capable of causing such infections, or if some of them have a heightened tendency to do so when they enter either the bloodstream or a deep tissue. FCAI scientists Johan Pensar and Jukka Corander joined a team of microbiologists and geneticists to unravel this mystery. By combining large-scale population genomics and in vitro measurements of immunologically relevant features of these bacteria, they were able to use machine learning to successfully predict the risk of developing a serious, and possibly life-threatening infection from the genomic features of a bacterial isolate. This opens the door for future technology where high-risk genotypes are identified proactively when a person is to undergo a surgical procedure, which has high potential to reduce the burden of nosocomial infections caused by S. epidermidis.
Research article: https://www.nature.com/articles/s41467-018-07368-7
Researchers are clearing language barriers for automated news, aiming for an increasingly varied view of the world
Researchers at the University of Helsinki are developing news automation and information retrieval from text masses in cooperation with five other universities and the Finnish News Agency STT.
How to automatically find the essential content of news in various languages? How might a computer produce news smoothly, and does technology adapt to small linguistic areas such as Finland?
These are among the challenges to be tackled by EMBEDDIA, a research project launching in 2019 with EU funding, with the University of Helsinki participating. The three-year project will be developing methods for automated text analysis and generation.
“Combining news written in several languages widens the perspective on the subject at hand, while making it possible to find out what is written on the item in different languages and in different media. The goal is to improve people’s access to information,” says Professor of Computer Science Hannu Toivonen, whose research group is taking part in the project.
Many media companies are already employing automated news for reporting on sports and elections. Using structured data, computers are able to write news articles. For example, ice hockey games are a comfortably regular phenomenon from a computational viewpoint: they consist of three periods, resulting in an unambiguous number of goals.
According to Toivonen, news automation is useful because it enables the production of a great amount of news from consistent data. Computers can write articles on local hockey games even for a handful of readers.
“In such cases, the audience of a single piece of news may be small, but when the number of articles is great, media businesses both achieve extensive coverage and respond to specific needs,” Toivonen explains.
For now, automated news is comprised of election and sports coverage, and the like, which is generated in a structured manner from structured data. In-depth profiles and news analyses produced by computers are still some way off in the future, since computers are yet unable to handle the linguistic and content variation of these text types.
FCAI Society has kicked off several initiatives for societally and ethically cognizant AI development
Set up in early 2018, FCAI Society has now brought together a multidisciplinary group of researchers and artists working on artificial intelligence and its wide impact in society.
In its latest meeting on 6 November 2018 FCAI Society discussed with representatives of the Ethics group of the Finnish government’s AI Programme (tekoalyaika.fi). The FCAI Society is willing to act as a discussion partner to the Programme in matters of AI ethics.
FCAI Society will make an inventory of existing ethical guidelines for AI, with the view to produce a set of guidelines to facilitate and promote ethical thinking in AI development. The guidelines will work not only as an internal guideline for FCAI’s operations and research, but also provide researchers and practitioners outside academia an example checklist for adopting and applying artificial intelligence tools and methods. FCAI is keen to take a facilitating role in the public discussion concerning ethics of AI, and in this context, the manifesto would serve as a common reference point to spark future dialogue on the subject.
The FCAI Society will be active in suggesting new research programs for the Academy of Finland and will also initiate joint research project proposals on the subject of AI and Society.
FCAI Society also has a series of podcasts planned concerning AI. The podcasts will feature FCAI Society members and other guests who will discuss the meaning, impact, hopes and risks related to artificial intelligence. The motivation behind the series is educational, both for the audience and the participants—a though-provoking chance to learn from different points of view how things like intelligence, privacy, art, work and creativity will be shaped by AI technologies.
FCAI scientists join forces with physicists to significantly improve the chances of detecting dark matter
Existence of dark matter has been inferred from gravitational observations in macroscopic scale surveys of the universe, but no dark matter has been directly detected yet. The elusive nature of dark matter has made its detection into one of the grand quests of modern physics. By joining forces with a Stockholm physics team led by professor Jan Conrad, FCAI scientists Umberto Simola and Jukka Corander have made a computational breakthrough which significantly improves the chances of directly detecting dark matter from a liquid xenon experiment currently conducted under the Gran Sasso mountain in Italy. The breakthrough is outlined in their scientific report submitted for publication and it is based on the AI-powered inference software platform ELFI, that is developed and maintained by a team of FCAI scientists. By exploiting the advanced features of ELFI, which uses smart machine learning techniques for deducing unknown functions by computer simulations, Simola and Corander were able to outperform all the currently existing methods for inferring dark matter particles, including those based on neural networks that have been popular for numerous AI applications. The team is excited by their breakthrough and is currently investigating possibilities for further advances in high-energy physics applications.
FCAI sets out to build AI toolbox
Finnish Center for Artificial Intelligence has received €1.4M funding from The Future Makers program of the Technology Industries of Finland Centennial Foundation and the Jane and Aatos Erkko Foundation. With the funding, FCAI will build an AI software toolbox to enable companies to have a smoother transition into using artificial intelligence methods.
Even though AI has been talked about the world over for quite some time now, its full potential still remains largely untapped. The development of new solutions is slowed down by a lack of top experts, for which there’s already a fierce global competition.
“We are designing software tools with which companies can develop the AI solutions they need—instead of building AI-assisted software tools from scratch. This means you can apply AI without having extensive in-depth knowledge of AI. Our overall goal is to enable the Finnish technology industry to retain control over the core AI technology they use,” says Samuel Kaski, Academy Professor at Aalto University.
HOPE project receives funding from the EU develop real-time monitoring of air quality
The European Union has conferred 4.7 Million euros to fund the production of real-time and detailed air-quality information using a 5G network. Some of the members in the HOPE project are particle scientists and computer scientists from the University of Helsinki.
A system for real-time, dependable, and detailed information on air quality is being developed for the residents of Helsinki. Three environments with different air qualities are being monitored within the HOPE project: Mäkelänkatu–Kumpulankatu, Pakila, and Jätkäsaari.
The University of Helsinki is developing the use of a 5G network within the HOPE project along with air-quality monitoring and a new air-quality index. The purpose of the new index is to take more variables than before into consideration, to gain more exact data on the real air quality.
“As cities grow more densely populated, the study of the air we breathe becomes increasingly important, since the deterioration of air quality is a growing concern to many. This project hopes to find tangible solutions to improving people’s wellbeing. One of these solutions is the new index for air quality,” says professor of aerosole and environmental physics Markku Kulmala.
To transfer the real-time data on air quality, 5G networks are being used, and one of the project’s objectives is to develop the 5G abilities of devices.
“The project enables both measuring devices that are fixed in the urban space and crowd-sourced devices, which are all utilised to compile a detailed image of the air quality. The goal is to summarise the main factors of air quality and health as clearly and exactly as possible,” says professor of computer science Sasu Tarkoma.
“The mapping of regional and temporal variations in air quality also gives us valuable information for studying the factors impacting it, while we can produce information about it to the public with the help of map applications,” says geo-informatics professor Petri Pellikka.
The part the University of Helsinki plays in the project belongs to the larger MegaSense research programme. The purpose of MegaSense is to create a world-wide observation system that will give us detailed information about air quality and harmful substances in the air. This information can be utilised in various devices in the car, at home, and at the office.
Other partners in the HOPE project are the City of Helsinki, Forum Virium Helsinki, Vaisala, HSY, Useless, and the Finnish Meteorological Institute. The project gained funding from the EU programme Urban Innovative Actions. The programme funds projects that look for innovative models to solve persistent urban problems and test them. The funding searches for projects that are both innovative and bold, and involve the community.
Drones in Urban Environment
Drones may change the urban landscape sooner and more dramatically than we might expect. In a matter of years, there may be thousands of drones flying over our heads every day. Business, science and security actors are fast developing new uses for drones. Drone visions and threats pose new challenges for lawmakers and public regulators. But combined with advancements in data science, AI and robotics, drones may be a significant growth driver in the near future.
The Drones in Urban Environment event, organised in Think Corner on 25 September 2018, covered these and many more issues related to unmanned aerial vehicles.
In the event, the keynotes and discussion were joined by leading figures from business, academia and the public sphere. Opening speech by Finnish prime minister Juha Sipilä was followed by the senior manager in strategy and innovation at Airbus, Vassilis Agouridas, who talked about development in smart mobility. Professor Alf Rehn shared insight on drone ambitions in emerging industries.
The panel discussion “1000 drones over Helsinki by 2020 – how?” covered the future developments in business and research, logistics, civil aviation, security and regulation, with top-ranking experts from these fields. Finally, the head of the drone capacity of the Finnish police force, Sami Hätönen, concluded the event with an exciting live demo of an urban drone search and surveillance mission.
A video recording of the event is available.
Drones are emerging as a promising new platform for research
The prospects related to unmanned aerial vehicles, also known as drones, range from parcel deliveries to self-guided smart devices that can, for example, search for criminal suspects. In common predictions, drones are considered helpful in controlling traffic, public events and riots.
“Drone technology provides new ways of monitoring, mapping and navigating in the urban environment for improving the safety, security and efficiency of cities and the wellbeing of citizens. Together with 5G, AI and modular sensing systems, it is possible to gather high quality and real-time data about cities,” says Sasu Tarkoma, professor of computer science at the University of Helsinki.
Another future possibility is to employ unmanned aircraft in the analysis of air quality in cities. Such use has been planned, for example, in the MegaSense project.
“We are currently investigating whether drones can be installed with air quality measuring equipment. Traditionally, air quality sensors are immobile or installed in cars, but drones could cover areas in a more versatile manner,” explains Petri Pellikka, professor of geoinformatics at the University of Helsinki.
Researchers are also able to gather information from locations that could otherwise only be accessed through hard work, lots of money and intricate scheduling. In addition to traditional cameras, drones can be installed with laser scanners or hyperspectral cameras, which are getting lighter every day. In addition, they are able to get closer to the ground than airplanes, which makes their images more accurate than those collected by planes. Such information can be utilised in geology and ecology, as well as agriculture.
“In the world-wide Shanghai Ranking, the University of Helsinki was this year placed 20th in the field of remote sensing. Remote sensing carried out with drones could take us even higher,” says Pellikka.
You can't tell whether an online restaurant review is fake - but this AI can
Researchers find that AI-generated reviews and comments pose a significant threat to consumers, but machine learning can help detect the fakes.
Sites like TripAdvisor, Yelp and Amazon display user reviews of products and services. Consumers take heed: nine out of ten people read these peer reviews and trust what they see. In fact, up to 40% of users decide to make a purchase based on only a couple of reviews, and great reviews make people spend 30% more on their purchases.
Yet not all reviews are legitimate. Fake reviews written by real people are already common on review sites, but the amount of fakes generated by machines is likely to increase substantially.
According to doctoral student Mika Juuti at Aalto University, fake reviews based on algorithms are nowadays easy, accurate and fast to generate. Most of the time, people are unable to tell the difference between genuine and machine-generated fake reviews.
“Misbehaving companies can either try to boost their sales by creating a positive brand image artificially or by generating fake negative reviews about a competitor. The motivation is, of course, money: online reviews are a big business for travel destinations, hotels, service providers and consumer products,” says Mika Juuti.
In 2017, researchers from the University of Chicago described a method for training a machine learning model, a deep neural network, using a dataset of three million real restaurant ratings on Yelp. After the training, the model generated fake restaurant reviews character by character.
There was a slight hiccup in the method, however; it had a hard time staying on topic. For a review of a Japanese restaurant in Las Vegas, the model could make references to an Italian restaurant in Baltimore. These kinds of errors are, of course, easily spotted by readers.
To help the review generator stay on the mark, Juuti and his team used a technique called neural machine translation to give the model a sense of context. Using a text sequence of “review rating, restaurant name, city, state, and food tags”, they started to obtain believable results.
“In the user study we conducted, we showed participants real reviews written by humans and fake machine-generated reviews and asked them to identify the fakes. Up to 60% of the fake reviews were mistakenly thought to be real,” says Juuti.
Juuti and his colleagues then devised a classifier that would be able to spot the fakes. The classifier turned out to perform well, particularly in cases where human evaluators had the most difficulties in telling whether a review is real or not.
The study was conducted in collaboration with Aalto University’s Secure Systems research group and researchers from Waseda University in Japan. It was presented at the 2018 European Symposium on Research in Computer Security in September.
Mika Juuti, Bo Sun, Tatsuya Mori, N. Asokan:
Stay On-Topic: Generating Context-specific Fake Restaurant Reviews.
Hate speech detecting AIs are fools for "love"
State-of-the-art detectors that screen out online hate speech can be easily duped by humans, shows a new study.
Hateful text and comments are an 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 University Secure Systems research group has discovered weaknesses in many machine learning detectors currently used to recognize and keep hate speech at bay.
Many popular social media and online platforms use hate speech detectors that a team of researchers led by Professor N. Asokan have now shown to be brittle and easy to deceive. Bad grammar and awkward spelling – intentional or not – might make toxic social media comments harder for AI detectors to spot.
The team 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, doctoral student at Aalto University.
Google Perspective ranks the “toxicity” of comments using text analysis methods. In 2017, researchers from the University of Washington showed that Google Perspective can be fooled by introducing simple typos. Gröndahl and his colleagues have now found that Perspective has since become resilient to simple typos yet can still be fooled by other modifications 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 stand-alone solutions.
The researchers recommend that more attention be paid to the quality of data sets used to train machine learning models—rather than refining the model design. The results indicate that character-based detection could be a viable way to improve current applications.
The study was carried out in collaboration with researchers from University of Padua in Italy. The results will be presented at the ACM AISec workshop in October.
Tommi Gröndahl, Luca Pajola, Mika Juuti, Mauro Conti, N.Asokan:
All You Need is “Love”: Evading Hate-speech Detection.
Elements of AI becomes the most popular course ever at the University of Helsinki
The Elements of AI MOOC organised by the University of Helsinki, FCAI and Reaktor awarded diplomas to the first graduates and received endorsement from the President of Finland in the graduation ceremony held on the 6th of September, 2018. With approximately 90 000 registered participants, it has become the most popular course ever at the University of Helsinki.
Antti Poikola joins the AI Accelerator
HIIT researcher Antti ”Jogi” Poikola has been selected to lead the AI Accelerator, one of the instruments realising the Finnish AI Strategy (www.tekoalyaika.fi), together with Alexander Törnroth. While working for the AI Accelerator Jogi will still continue part-time working at Aalto University completing his PhD thesis. Jogi is founder of the Open Knowledge Finland association and a leading figure in the Internationally growing MyData movement. He is one of the main organizers of the MyData conference series.
Jogi says that the AI Accelerator is not for businesses taking the first steps towards utilising AI, but for companies already more advanced. They can then mentor each other and join forces in particular topics of common interest. Jogi’s example is chatbots in Finnish; many companies in different industrial segments would need solutions handling well Finnish text. The idea of the AI Accelerator is really to “be more by working together”. The accelerator process will be designed during this autumn. Companies need to be encouraged to think more boldly. The AI Accelerator is aimed at companies utilising AI, not directly for companies providing AI solutions, although some matchmaking can be foreseen between utilisers and providers. Jogi expects that some tens of companies will participate in the AI Accelerator. The AI Accelerator is working under the auspices of Technology Industries of Finland, an association represented in the HIIT Board by director Mervi Karikorpi (deputy: director Ville Peltola).
Make-up of superbug MRSA revealed – with prospective methods to prevent inter-species transfer
An international team of researchers, including Professor Jukka Corander (University of Oslo, University of Helsinki), has mapped the entire genetic make-up of over 800 strains of the common superbug MRSA, or Methicillin-resistant Staphylococcus aureus. The bacteria is known best for its world-wide prevalence in hospital environments.
Superbugs like MRSA are resistant to most antibiotics and can lead to life-threatening or deadly infections in humans. MRSA is common also in live stock and causes, for instance, mastitis in cows and skeletal infections in chickens.
According to the study, humans are the most likely original carrier of the bacteria, but the source for the current strains infecting humans are cows. The researchers now understand the mechanisms of how the bacteria is able to transfer from one species to another thanks to a thorough understanding of its genome. When jumping species, the bacteria is able to acquire new genes that help it thrive in the new environment.
Detailed analysis of the changes in the genetic make-up of the bacteria achieved now could offer a way to develop new anti-bacterial treatments. Knowledge of the transmission can also help devising strategies to prevent the bacteria from developing antibiotic resistance, or to block its access to humans altogether.
The results have been published in Nature Ecology & Evolution.
Link to the article: nature.com/articles/s41559-018-0617-0
Read more on Sanger Institute website: sanger.ac.uk/news/view/gene-study-pinpoints-superbug-link-between-people-and-animals
A new computational method for reconstructing the evolution of a tumor
Discovering the evolution of a tumor may help identify driver mutations and provide a more comprehensive view on the history of the tumor. One way to tackle this problem is to take several samples from the tumor, find the mutations they contain, and then determine a plausible evolution of the tumor that gave rise to these mutations.
A multi-disciplinary team lead by University Researcher Alexandru Tomescu from the University of Helsinki and Prof. Martin Milanič from the University of Primorska (Slovenia) has developed a new method for discovering the evolution of the tumor, using DNA sequencing data from multiple samples of a tumor. This method proved more accurate than the existing methods, both on synthetic data and on data from several types for cancer, such as clear cell renal cell carcinoma, high-grade serous ovarian cancer, breast cancer xenoengraftment and uterine leiomyomas.
This research has been published in the journal Bioinformatics, and is available under an Open Access license: Edin Husić, Xinyue Li, Ademir Hujdurović, Miika Mehine, Romeo Rizzi, Veli Mäkinen, Martin Milanič and Alexandru I. Tomescu, MIPUP: Minimum perfect unmixed phylogenies for multi-sampled tumors via branchings and ILP.
This computational method is based on a previous theoretical result proving a relation between this bioinformatics problem and a problem on directed graphs, presented at the WG 2017 graph-theory conference (Ademir Hujdurović, Edin Husić, Martin Milanič, Romeo Rizzi, Alexandru I. Tomescu, The Minimum Conflict-Free Row Split Problem Revisited. WG 2017: 303-315) and published in the journal ACM Transactions on Algorithms (Ademir Hujdurovic, Edin Husic, Martin Milanic, Romeo Rizzi, Alexandru I. Tomescu, Perfect Phylogenies via Branchings in Acyclic Digraphs and a Generalization of Dilworth’s Theorem. ACM Trans. Algorithms 14(2): 20:1-20:26).
An example is shown in the figure above: Tumors usually have a tree-like evolution, with new mutations (c1, c2, c3, …) accumulating on each edge of the tree. DNA sequencing samples (r1, r2, r3, r4) usually mix several leaves (A, B, C, D, E) of the tree, making it hard to discover the actual evolution.
University of Helsinki brings together a corporate cluster to develop techniques for measuring air quality
The University of Helsinki will start cooperating with Finnish companies within a Business Finland project to create a consortium with the purpose of developing businesses based on the MegaSense programme for improving the availability of air-quality data.
“MegaSense has every opportunity to reach significant scientific breakthroughs and a wide deployment of the results”, says the head of the project, Professor Sasu Tarkoma of the Department of Computer Science at the University of Helsinki.
The companies involved are experts in sensor technology, mobile networks, and data management. The shared goal of the university and the companies is to make a hundred million euros in the years 2020–2025 on products and services developed around air-quality measurements.
MegaSense brings together expertise in atmospheric sciences, computer science, and geo-informatics at the University of Helsinki. The idea of MegaSense is to combine a large number of low-cost sensors measuring outdoor air quality with extremely precise SMEAR concept stations to form a network producing data in real-time. Citizens can follow and monitor the air quality in their streets and across large areas, such as cities.
Password managers vulnerable to insider hacking
Researchers from Aalto University and the University of Helsinki have found over ten computer security-critical applications that are vulnerable to insider attacks. Most of the vulnerabilities were found in password managers used by millions of people to store their login credentials. Several other applications were found to be similarly susceptible to attacks and breaches across the Windows, macOS and Linux operating systems.
Computer software often starts multiple processes to perform different tasks. For example, a password manager typically has two parts: a password vault and an extension to an internet browser, which both run as separate processes on the same computer.
To exchange data, these processes use a mechanism called inter-process communication (IPC), which remains within the confines of the computer and does not send information to an outside network. For this reason, IPC has traditionally been considered secure. However, the software needs to protect its internal communication from other processes running on the same computer. Otherwise, malicious processes started by other users could access the data in the IPC communication channel.
"Many security-critical applications, including several password managers, do not properly protect the IPC channel. This means that other users’ processes running on a shared computer may access the communication channel and potentially steal users’ credentials," explains Thanh Bui, a doctoral candidate at Aalto University.
While PCs are often thought to be personal, it is not uncommon that several people have access to the same machine. Large companies typically have a centralised identity and access management system that allows employees to log into any company computer. In these scenarios, it is possible for anyone in the company to launch attacks. An attacker can also log in to the computer as a guest or connect remotely, if these features are enabled.
"The number of vulnerable applications shows that software developers often overlook the security problems related to inter-process communication. Developers may not understand the security properties of different IPC methods, or they place too much trust in software and applications that run locally. Both explanations are worrisome," says Markku Antikainen, a post-doctoral researcher at the University of Helsinki.
Following responsible disclosure, the researchers have reported the detected vulnerabilities to the respective vendors, which have taken steps to prevent the attacks. The research was done partly in co-operation with F-Secure, a Finnish cyber-security company.
Prospects and challenges of AI discussed in a conference on EU innovation strategy
Artificial intelligence was one of the main topics in the recent conference Innovative Europe? Time for a new EU innovation strategy. In a workshop focusing on the impact of AI on working life, HIIT Director Petri Myllymäki brought up some of the shortcomings of the current generation of AI systems and stated that the next wave of AI will be more transparent, will be able to explain itself and better assess uncertainties.
These are some of the core principles guiding the world-class research done at the Finnish Center for Artificial Intelligence. "The next generation will be the big revolution. We haven't seen anything yet," Myllymäki said. The technology is still in its formative stages, and Europe has a good chance of catching up on the early American and Chinese lead.
Researchers propose new solutions to alleviate hostility on social media
Disputes could be alleviated in arenas where people voicing incendiary opinions would have to deal with the emotions and reactions of a variety of readers, including their own near and dears.
Arguments and the echo chambers recycling strong opinions on social media are actively studied all over the world. Many researchers have suggested that inciting hatred and hostility between extreme opposites on an issue could be alleviated by wide-ranging content recommendation across hard-set divides. This way, people could be exposed to content and perspectives that differ from their own fixed opinions.
A joint research effort of Aalto University, University of Helsinki and Syracuse University challenges the approach that has attracted a great deal of attention. The researchers analysed approximately 50 000 messages from a total of ten Finnish Facebook groups, half of which had a positive stance towards immigration and the other half were strongly against it. Only two per cent of the links shared in the opposed groups were the same; consequently, the groups do not consume the same content, save only a few exceptions.
After this network analysis, they took a more detailed look at approximately 1 000 comments submitted to the few pieces of content that were shared on both sides. The qualitative analysis showed that Facebook as a platform does not encourage users to encounter others and their opinions, even though in principle it provides an arena for open debate.
"The original content may be a neutral piece of news, but the person sharing it sets it against their own opinion. At the same time, those who disagree may be negatively labelled – and the framing steers the debate onwards. This happens equally in groups that are in favour of immigration and against it. Therefore, we do not believe that content recommendation as such can be the solution, but we need new approaches," says Salla-Maaria Laaksonen, researcher at the University of Helsinki.
Based on the research carried out in the field, the team have come up with new solutions to alleviate hostile echo chambers.
"We could create platforms and groups that highlight the importance of listening and reflecting upon the opinions of others. The person expressing an opinion would then be forced to face the readers' emotional reactions. Instead of judging others, people sharing content would have the opportunity to contemplate what they have said and how it affects others," Aalto University researcher Matti Nelimarkka explains.
"One possible factor explaining and intensifying the polarisation of opinions and attacks against other individuals is the circle of silence. It is created when, for example, a person in favour of immigration is afraid to discuss the possible effects that reception centres may have on communities that host them. The debate is bigoted and aggressive on both sides to such an extent that participants are afraid to voice differing opinions," says Nelimarkka.
Researchers are also looking into the possibility of using the social relationships of individuals in content recommendation. For example, an algorithm could recommend a discussion in which a relative or a friend is involved. A person sharing hostile or bigoted content would then be exposed to differing perceptions of people they know, and the dispute would no longer take place just between strangers.
"Our ideas still require further development so that we can guarantee that the recommendation algorithm does inflict any harm to people’s relationships and that the algorithm can recommend mind-opening content regardless of where a person might stand on a given issue," Nelimarkka notes.
Research article: https://dl.acm.org/citation.cfm?id=3196764
VTT joins FCAI as third founding member
Technical Research Centre of Finland VTT will join Finnish Center for Artificial Intelligence FCAI launched by Aalto University and the University of Helsinki as a third founding member.
VTT will bring their strong industry networks and their know-how in applied technology to the FCAI community. Their help will enforce FCAI's ability to put the top research in both founding universities into far-ranging and efficient use in companies, public organisations and society at large.
FCAI promotes high-quality research and education on artificial intelligence in Finland and the applicability of AI to benefit companies and society. VTT will expand FCAI's ability to speed up the necessary renewal and competitiveness of Finnish industry through AI-based innovations.
FCAI strives to make the new generation of AI methods a reality: create AIs that are understandable, trustworthy, and data-efficient. FCAI's goal is to expand into a national network of universities, companies and research institutions who will lay the groundwork for Finland to become a global leader in AI research and AI applications.
Growth in any strand of industry depends on the ability to make use of cutting-edge technology. Artificial intelligence is the key leverage here.
"Our vision is to bring our high-class research in several strands of artificial intelligence to benefit people's every-day lives, companies and public bodies. FCAI's impact is a potent mixture of research, a network of startups, doctoral education and competence building in AI, new innovative products and services, and smart experiments in public administration," says Director of FCAI, Academy Professor Samuel Kaski.
"The single most significant growth factor now is applying artificial intelligence and ICT in general. For citizens, new innovations and solutions will bring a change in work content, professional skills, and the services society provides. AI will be able to make, for instance, medical care more efficient and personalised," says Tua Huomo, Executive Vice President at VTT.
FCAI is building a national hub of universities, research institutes, industry and the private sector and public organisations with strong international networks. The FCAI community is constantly expanding with new memberships and projects.
A new computational tool for determining the composition of bacterial communities
Determining the composition of bacterial communities at strain level resolution is critical for many applications in infectious disease epidemiology and in bacterial ecology. Using the latest advances in computational inference and sequence analysis, an international team led by professors Jukka Corander and Antti Honkela has developed a new metagenomics tool called mSWEEP, which goes significantly beyond the state of the art in this field. The effectiveness of mSWEEP is demonstrated with infection data from major human pathogens and it is expected to pave the way for entirely new approaches to addressing important biological and clinical questions about inter-strain competition, dissemination of resistance and virulence.
A preprint of an article describing mSWEEP is available from bioRxiv.
Espoo becomes a member of FCAI
The City of Espoo has become a member of the Finnish Center for Artificial Intelligence FCAI. FCAI is a research centre launched by Aalto University and University of Helsinki, which gathers together the best artificial intelligence researchers in Finland. FCAI's objective is to make the most advanced methods of artificial intelligence available to enterprises, organisations and society.
The City of Espoo sees that developing artificial intelligence together will be beneficial for the whole innovation community from enterprises to R&D organisations and the inhabitants in Espoo.
"For a researcher, the data in the databases of the city of Espoo and the shared databases of the Helsinki metropolitan area is very interesting. Especially the innovative start-up companies in the area and Espoo's desire to be profiled as a pioneer in the use of intelligent technologies set a good basis for cooperation with researchers developing artificial intelligence. We have all the prerequisites to expand our cooperation to other research centres and other cities as well," says the Director of FCAI, Academy Professor Samuel Kaski.
"On the one hand, researchers need data for the development of artificial intelligence methods and technology, and public organisations have this data. On the other hand, we as a city get to use the methods, technologies and the latest knowledge of artificial intelligence research in the development of our services," says Tomas Lehtinen, data analyst consultant for the City of Espoo.
Making efficient use of sensitive big data and keep it safe and private?
A new method developed by FCAI researchers of University of Helsinki and Aalto University together with Waseda University of Tokyo can use, for example, data distributed on cell phones while guaranteeing data subject privacy.
Modern AI is based on learning from data, and in many applications using data of health and behaviour the data are private and need protection.
Based on the concept of differential privacy, the method guarantees that the published model or result can reveal only limited information on each data subject while avoiding the risks inherent in centralised data.
The digital newseye reads stories about past occurrences and even explains them to new readers
NewsEye, the multidisciplinary research project at the University of Helsinki, is developing a research assistant automated by artificial intelligence so that experts in digital humanities, computer science, and library science can work on digitalised memory material. They wanted to cooperate in order to gain results that are relevant and usable for research, teaching, and other uses of the material.
The focus of the research is on data science, and the project is headed by Professor Hannu Toivonen.
"The most interesting research object is the automated research assistant that can use new tools developed in the project independently to search for results that are interesting to the user, report its findings in clear text, and can explain the findings and its own work. This is our objective in Helsinki," Professor Toivonen, known as a specialist in creative computing, says.
The National Library will deliver historical Finnish newspaper material from the years 1771-1910 for NewsEye to process. It has digitalised all the Finnish newspapers that appeared during this time period and made them into a machine-readable data packet. The material will be complemented with newspaper material from 1911-1917.
The project is also working on enriching text automatically by recovering names and attitudes from text. The Finnish researchers are also focusing on developing new tools for analysing enriched text from different viewpoints so that different contexts and baselines are observed.
From the University of Helsinki, the participants are Professor Hannu Toivonen, historian Mikko Tolonen and his research group, and from the National Library, Minna Kaukonen and her group of researchers. Similar tri-disciplinary teams from France and Austria are participating, as well as one German partner.
The multi-lingual feature is a novelty; the methods and tools will be made as independent of language as possible, or at best they will be able to work with different languages at the same time. According to the researchers, this is important - but very rare - in a European context.
AI-created family trees confirm class divisions in Finland in the 18th and 19th century
The genealogy algorithm AncestryAI efficiently combines huge amounts of birth data.
It would take 100 person-years for a genealogist to map and find all the parents for five million people – with a rate of one person per minute. The AncestryAI algorithm can do the same work in an hour using 50 parallel computers and with a success rate of 65 percent. The algorithm can also measure the level of uncertainty for each connection so that unreliable results can be ignored.
"The algorithm does not replace the work of genealogists; it is simply a tool for helping them in their work. The genealogy algorithm can suggest connections which are probably correct, but on its own it is not as precise as a careful genealogist. The algorithm can also search for parents from nation-wide data, while a genealogist may need to limit their search to just one parish," explains Eric Malmi, doctoral student at Aalto University who currently works for Google in Zürich.
Malmi will defend his doctoral dissertation at Aalto University in June in the supervision of Aalto University professor and FCAI programme leader Aristides Gionis.
Are social media echo chambers?
Situations where one is exposed only to opinions that agree with their own are an increasing concern for the political discourse in many democratic countries.
In a recent study, researchers from the University of Helsinki and Aalto University compared the political leaning of individual users with the political leaning of their connections on Twitter. To do that, they define two numerical scores: a production score, which reflects the political leaning, liberal or conservative, of the messages published, or ‘produced’, by a user; and a consumption score, which reflects the political leaning of the messages received, or ‘consumed’, by a social-media user.
“By comparing the two scores for each user, we find that Twitter users are, to a large degree, exposed to political opinions that agree with their own,” says Michael Mathioudakis, Assistant Professor at the Department of Computer Science from the University of Helsinki, and one of the authors of the study.
The researchers also find that users who post messages with diverse leaning pay a 'price of bipartisanship' in terms of their network centrality and content appreciation.
“That is, the number of times their messages, tweets, are shared or rated positively, ‘favorited’, by other users is smaller than that of partisan users,” says Mathioudakis.
Even as the researchers found political discussions on Twitter that resemble echo chambers, they also identified ‘gatekeepers’ - a class of users who consume content with diverse leaning but produce partisan content with a single-sided leaning.
“Such users might play an important role in enabling the flow of information between the two sides, but this is something that we’ll look more into in future research”, he says.
The researchers performed their analysis on small and larger sets of tweets. The smaller sets included tweets that spanned one week and were focused on a specific discussion, for example about Obamacare during June 22-29 in 2015. The largest set consisted of 2.6 billion tweets generated by users who were active in discussions around US politics, from 2009 to 2016. The large dataset was obtained from previous work by Dr Kiran Garimella from Aalto University, a co-author here.
The traits of fast typists discovered by analysing 136 million keystrokes
An online study with 168,000 people shows large variation in typing speeds and styles.
Researchers from Aalto University in Finland and University of Cambridge in the United Kingdom collected extensive data about the typing behaviour of 168,000 volunteers. The researchers developed an online typing test following scientific standards and published it on the free typing speed assessment website typingmaster.com.
"Ethical large-scale crowdsourcing experiments that allow us to analyse how people interact with computers on a large scale are instrumental for identifying solution principles for the design of next-generation user interfaces," says Dr Per Ola Kristensson, University Reader (Associate Professor) in Interactive Systems Engineering at the University of Cambridge.
The dataset is the largest ever on everyday typing and exposed several factors that differentiate fast vs. slow typists. In addition to making less errors, the researchers found that fastest typists rely on so-called "rollover" where a letter key is typed already before the previous one is released. “This strategy is only possible for highly practiced letter combinations and when performance does not rely on visual attention,” says Anna Feit. “An important goal for typing is to learn not to look at fingers.”
The results of the study change our understanding of typing performance. Most of our knowledge of how people type is based on studies from the typewriter era. People presently make different types of mistakes: more errors where a letter is replaced by another one, whereas in the typewriter era they often added characters or omitted one. Also, modern users use their hands differently. Anna Feit explains: “Modern keyboards allow typing keys with different fingers of the same hand with much less force than what was possible with typewriters. This is partially explaining why self-taught typing techniques using less than ten fingers can be as fast as the touch typing system, which was probably not the case in the typewriter era.”
Analysing the individual keypresses, the researchers found that users exhibit different typing styles, characterised by how they use their hands and fingers, the use of rollover, tapping speeds, and typing accuracy. It is now possible for a computer to classify users' typing behaviour simply based on the observed keystroke timings which does not require to store the text that users have typed. Such information can be useful for example for spell checkers, or to create new personalised training programmes for typing.
The study will be presented at the world's largest computer-human interaction conference, CHI, at Montreal, Canada, USA, in April 2018, where it was recognised to be among the top 5% of publications. The study was done in collaboration with TypingMaster.com.
The anonymised dataset is published and free to use for research purposes.
AI improves touchscreen interfaces for users with impairments
A new AI method adapts touchscreen interfaces to make more out of the capabilities of ageing users and users with disabilities.
Researchers at Aalto University and FCAI, Finland, and Kochi University of Technology, Japan, developed a new algorithmic approach to user interface optimization that takes individual differences into account.
”The majority of available user interfaces are targeted at average users. This “one size fits all” thinking does not consider individual differences in abilities – the ageing and disabled users have a lot of problems with daily technology use, and often these are very specific to their abilities and the circumstances,” says postdoctoral researcher Jussi Jokinen at Aalto University.
”There are ways to automatically optimize the user interface, but this is efficient only if we have a realistic model of the user. Previously, designers did not have detailed models that are based on psychological research and can be used to predict, how different individuals perform in interactive tasks."
FCAI Society: understanding and communicating AI across scientific divides
Solving the major technical hurdles in artificial intelligence, FCAI has now brought together the top expertise in both Aalto University and University of Helsinki in the technical development of AI.
However, we still need a holistic view and understanding of artificial intelligence across scientific borders in order to also engage the public in the changes AI will bring.
FCAI has sought experts from philosophy, ethics, sociology, legal studies, psychology and art to explore the impact AI will have in all aspects of our lives.
This cross-disciplinary group, FCAI Society, will in interaction with FCAI researchers consider the wide implications of AI research and furthermore the FCAI Society and FCAI researchers will together engage in public dialogue.
FCAI Society has teamed up with the event venue Think Corner at the University of Helsinki to expose AI research to public interest and scrutiny in an ongoing series of themed events: debates, discussions and demos.
FCAI Society will try to meet the pressing need to engage in dialogue and bridge scientific divides. We will deepen understanding on both sides: both of what is technically possible and how AI methods affect societal change and global equality. The lessons we have to teach each other we will take with us to the public domain and engage everyone in improving our common AI literacy. Here Think Corner’s events, which consistently reach hundreds of people in their prime location in the Helsinki city center and many more online, will have a prominent role.
The group will not remain fixed but expand and change according to the goals, research interests and ongoing projects within FCAI. This way we can have insight into the ways AI methods will live on and be taken up different societal settings. FCAI Society will also remain open to future research collaborations.
The initial composition of the FCAI Society, subject to change:
- Hanna Haaslahti – artist
- Raul Hakli – university researcher, ethics (University of Helsinki)
- Sara Heinämaa – professor, philosophy (University of Jyväskylä)
- Timo Honkela – professor, language technology, philosophy of AI (University of Helsinki)
- Minna Huotilainen – principal investigator, cognitive science (University of Helsinki)
- Riikka Koulu – assistant professor, legal studies (University of Helsinki)
- Jaakko Kuorikoski – associate professor, philosophy (University of Tampere)
- Krista Lagus – professor, digital social science (University of Helsinki)
- Arto Laitinen – professor, philosophy (University of Tampere)
- Turo-Kimmo Lehtonen – professor, sociology (University of Tampere)
- Pekka Mäkelä – coordinator, ethics (University of Helsinki)
- Kasperi Mäki-Reinikka – artist
- Göte Nyman – professor emeritus, psychology (University of Helsinki)
- Mika Pantzar – professor, consumer research (University of Helsinki)
- Osmo Soininvaara – statistician, former government minister and member of parliament, Helsinki city council member
- Petri Ylikoski – professor, science and technology studies (University of Helsinki)
Pressing a button is more challenging than appears
Pressing a button appears easy, but the brain needs a probabilistic internal model to control a press. The action appears effortless and one easily dismisses how challenging it is. Researchers at Aalto University, Finland, and KAIST, South Korea, created detailed simulations of button-pressing with the goal of producing human-like presses.
The researchers argue that the key capability of the brain is a probabilistic model: the brain learns a model that allows it to predict a suitable motor command for a button. If a press fails, it can pick a very good alternative and try it out.
"This research was triggered by admiration of our remarkable capability to adapt button-pressing", tells professor Antti Oulasvirta of Aalto University. "We push a button on a remote controller differently than a piano key. The press of a skilled user is surprisingly elegant when looked at terms of timing, reliability, and energy use. We successfully press buttons without ever knowing the inner workings of a button. It is essentially a black box to our motor system. On the other hand, we also fail to activate buttons, and some buttons are known to be worse than others."
Interactive AI to address gestational diabetes
A new HIIT project co-funded by Business Finland has started where machine learning and HCI researchers are contributing to address gestational diabetes. The project is part of the Interactive AI research area of FCAI.
The project is originated from the CleverHealth Network, an ecosystem coordinated by the Hospital District of Helsinki and Uusimaa (HUS). The main partners in the gestational diabetes project are HUS, Aalto University, the University of Helsinki, Elisa and Fujitsu.
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.
"By improving lifestyle during pregnancy, we can probably reduce the number of mothers who will develop type 2 diabetes as well as the health risks to the child, thereby also improving the health of future generations. The application will help the patient to learn how her diet, activity and sleep affect blood glucose levels and weight gain and, consequently, the course of the pregnancy and the newborn’s health," says Saila Koivusalo, research director of the project and specialist in obstetrics and gynaecology.
The project will make use of machine learning to provide guidance and treatment that are in line with the patient’s risk profile and meet her individual needs. Artificial intelligence also makes it possible to draw up predictions of both the mother’s and the child’s future health.
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 professor Hannu Toivonen, tackles 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.
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.
Women in Data Science brings a global phenomenon to Helsinki
The Global Women in Data Science (WiDS) Conference aims to inspire and educate data scientists worldwide, regardless of gender, and support women in the field. This annual one-day technical conference provides an opportunity to hear about the latest data science related research and applications in a broad set of domains.
Since the inaugural conference in 2015, WiDS has gone global with over 80 regional events worldwide from 30 countries and more than 75,000 people from 75 countries participated in 2017.
WiDS Helsinki is supported by FCAI and HiDATA.