WASP Professor Proutiere’s work among top 30 at world’s largest AI conference

December 3rd marked the start of Neural Information Processing Systems (NeurIPS), the world’s largest and most prestigious conference in Artificial Intelligence (AI) and Machine Learning (ML), an arena to which merely the best are invited to present their research. This year one of those is Alexandre Proutiere, Professor in Automatic Control.

By: Louise Gustafsson

 

AI is one of the hottest topics of our time. It sparks interest in all parts of society, from the everyday man to industry and academia. A clear sign of its relevance is the rapid development of NeurIPS, the AI arena which has grown from 1200 to 10 000 visitors in a short period of time. This year tickets sold out in 10 minutes, almost as fast as to concerts by famous artists.

Fotograf: Danica Kragic https://www.kth.se/profile/dani/

“The conference is of tremendous importance for researchers in order get exposed to the most recent and promising technical and practical progresses made in AI, as well as for companies willing to explore and develop AI and ML applications to their respective businesses”, says Alexandre Proutiere, Professor at the Division of Automatic Control.

A total 1011 papers are published at NeurIPS in 2018, but the honor of presenting is only given to the 30 best. One of those presentations will be given by Proutiere and his research team.

“To have an oral presentation means of lot, both to my team and for KTH, as the corresponding paper will be highly visible during the conference”, Proutiere says.

The Department of Automatic Control has published papers at NeurIPS consistently for six years, an achievement that is unique in Sweden. KTH in general did well, setting a record with five accepted papers. Three of those papers comes from the Department of Automatic Control.

“It is critical in terms of visibility. It ensures that your ideas are going to be seen and guarantee a maximal impact of your research. Being consistently present and visible at the conference positions the university in the AI academic and industrial landscape. It is also a strong factor for attracting the best students and researchers”, Proutiere says.

The paper by his team is called “Exploration in structured RL”. RL, meaning Reinforcement Learning, is an increasingly important and popular branch in AI and ML with applications in fields such as games, finance and autonomous vehicles.

“Recently, RL algorithms have been combined to Deep Learning, with spectacular practical successes. Over the past three years, researchers at Deepmind have been able to design software that is able to learn to play video games better than humans, and to beat the world champion in the game of Go”, Proutiere says. “As of now, we are not able to theoretically understand the reasons of these successes. The gap between practical algorithms and theory has never been as wide as today. The paper we present at NeurIPS constitutes a first step towards filling up this gap”.

 

About NeurIPS

NeurIPS stands for Neural Information Processing Systems. NeurIPS is the largest, most influential and prestigious conference in Artificial Intelligence (AI) and Machine Learning (ML) and has been held every year in December since 1986. Montreal, Canada, is hosting the event this year.

The conference is an arena for researchers in academic and industrial to share their work and learn about the latest advanced techniques used in AI and ML.