The organization has been set up, the research programme established and PhD students appointed: the new AI4Science Laboratory is ready for take-off. A virtual kick-off workshop on Wednesday 8 July celebrates the appointment of five PhD students and the lab now being fully operational. Scientific director Dr Bernd Ensing expects AI4Science to break ground in data analysis and assist scientists in obtaining meaningful experimental results. He invites researchers to participate in the kick-off workshop.
Established in 2019 at the Faculty of Science, the AI4Science Lab aims to solve scientific data problems with modern machine learning approaches. It sets out to discover patterns in data streams from experiments in a wide variety of scientific fields, ranging from ecology to molecular biology and from chemistry to astrophysics. The lab is a joint initiative of the institutes for astronomy (API), biology (IBED), chemistry (HIMS), informatics (IvI), life sciences (SILS) and physics (IoP), and it is connected to AMLAB, the Amsterdam Machine Learning Lab.
Housed at the Informatics Institute (IvI), Dr Patrick Forré manages the lab on a day-to-day basis. Scientific director is Dr Bernd Ensing, associate professor of molecular simulation of materials and chemistry at the Van 't Hoff Institute for Molecular Sciences (HIMS).
As a chemist, Ensing knows the challenge of designing experiments to obtain meaningful data, and the struggle often associated with the interpretation of results. In fact, he says, this is a common aspect of scientific research that is at the heart of the effort of the AI4Science Laboratory. "A wealth of data is continuously produced from all sorts of experiments in the Faculty of Science research labs. The underlying research question we aim to answer is basically always the same: How can we detect, classify, and predict relevant patterns in scientific data if they are hidden within large amounts of non-relevant data?"
Thanks to automatization, parallelisation, high-throughput setups, high-resolution instruments, and fast networks, “big data” has become a practical issue in a wide range of experimental research projects. Analysis of these large data streams is often a grand challenge. The current surge of AI machinery provides a compelling opportunity to assist in the analysis of these scientific data streams. "At the AI4Science Laboratory, we develop and apply artificial intelligence and data-driven solutions for scientific discovery in the broadest sense", says Ensing.
To pursue this, the Lab has appointed five PhD students from different educational backgrounds on a range of interdisciplinary topics. Some of the students have a clear AI profile, while others are rooted in the traditional scientific disciplines.
All of them will develop and apply modern AI and Machine Learning techniques to tackle problems from completely different fields: predicting bird migration from radio data (Fiona Lippert, IBED), enhancing chemical discovery procedures (Jim Boelrijk, HIMS), interpreting signals from gravitational waves (Benjamin Miller, IoP), classifying space radio phenomena (David Ruhe, API) and unravelling causal relations in gene regulation networks (Teodora Pandeva, SILS).
To Ensing, the multidisciplinary nature of the AI4Science Lab is very exciting. "It's all about building bridges between computer and data science and the different experimental research fields", he says. "Our Lab has already become the centre of a rapidly growing consortium of people interested in machine learning for scientific discovery. We all exchange ideas through the biweekly AI4Science colloquium series and social media. So apart from the scientific breakthroughs we hope to achieve inside our Lab, I expect that in the coming years we will see a myriad of spin-off projects and collaborations, within UvA's Faculty of Science and outside.”
At the virtual kick-off symposium on Wednesday 8 July, all five PhD students of the AI4Science Lab will briefly highlight the current state-of-the-art and their future perspective on using AI in their respective scientific domains. Adding to this, several invited speakers will talk about intriguing scientific challenges of combining artificial intelligence techniques with working at the forefronts of scientific fields such as systems biology, particle physics, molecular modelling, and astrophysics. Registration is free for everyone with an interest in current developments in artificial intelligence for the natural sciences. Ensing invites all FNWI scientists wondering how machine learning could help their research to present a poster with their work and seize the opportunity for discussions with computer scientists.