Last year’s event was a great success with the official opening of dScience, and we plan this year’s event to be at least as good. This is an opportunity for Master’s students, PhD candidates and researchers to meet and discuss with each other and with people from the large data science community.
There will be interesting talks from three invited speakers with wide experience from the fields of data science and machine learning/AI. They are Morgan Briggs (Alan Turing Institute), Vegard Antun (UiO) and Pelle Sillrén (Spotify), working on current topics in AI and ethics, deep learning, data analysis and causality. In addition to the talks, a long range of companies have signed up to have stands, and we will serve food and soft drinks for free.
Morgan Briggs, who is a Research Associate for Data Science and Ethics at The Alan Turing Institute in London, will give the talk “AI Explainability from the ground up”. Some of her research interests in social data science are food security and remote sensing, renewable energy, and ethical considerations of digital technologies. In her talk, Briggs will explore the topic of AI Explainability, building upon explaining decisions made with AI, the co-badged guidance produced by the Information Commissioner’s Office and The Alan Turing Institute. The aim of the guidance is to help organizations in explaining the process, services and decisions delivered or assisted by AI, to the individuals affected by them. It currently serves as the most comprehensive practical guidance on AI explanation anywhere to date.
“Can all AI systems be trained?” is the second talk on the program, by Vegard Antun, who is a Postdoctoral Fellow at the Department of Mathematics, University of Oslo. Much of his research is concentrated on the stability and accuracy of methods for inverse problems. Machine learning methods and AI have proven to be superior in solving many computational problems, but there is also overwhelming empirical evidence that much of modern machine learning is non-robust (unstable) and may produce non-trustworthy results. In fact, modern algorithms do not seem to achieve the degree of robustness as established theoretical results suggest. In his talk, Antun will address this concern, and explain how there are fundamental limits on the AI systems one can build in practice.
From Spotify, we have invited the Senior Data Scientist Pelle Sillrén. Today, Spotify is the most popular music streaming service in the world with 433 million users. Sillrén will give the talk “Data informed product development using randomised controlled trials and other causal inference methods”. Here, you will learn about how Spotify uses large scale Randomised Controlled Trials to make data informed decisions that drive product development, and what to do when randomisation is not an option.
This year a large number of companies have signed up for stands. These include ABB, Abelee, Computas, DNV, Equinor, FINN.no, Gjensidige, Kongsberg Maritime, Microsoft, Norsk Regnesentral, NRK, Oda, Schibsted, Schlumberger, SINTEF, Soundsensing and Statens Kartverk. We highly encourage everyone attending to visit the stands to get to know more about these companies.