Research

RL Theory Seminars: bringing theory to the forefront


Amii Fellow and Canada CIFAR AI Chair Csaba Szepesvári has teamed up with Gergely Neu and Ciara Pike-Burke, researchers at the Pompeu Fabra University (Barcelona), to launch the RL Theory Seminar series. This series of online talks is geared at increasing knowledge-sharing in the rapidly growing field of reinforcement learning.

In recent years, reinforcement learning (RL) has grown to be one of the most active research areas in machine learning. Recent empirical successes have triggered a new wave of theoretical research in RL, and with so many new directions opening, it has become challenging to keep up on new developments. In response to this challenge, Szepesvári and colleagues launched the RL Theory Seminars.

“As someone who is keen on advancing the field, I’m thrilled to see the growing excitement around reinforcement learning,” says Szepesvári. “AI researchers around the world are taking notice of what we in the RL community have known all along – that these theories will be essential if we hope to tackle a large number of practical applications, ranging from problems in artificial intelligence to operations research or control engineering. RL theory provides the foundations we need to think through these problems.”

Amii CEO Cam Linke, also an RL researcher, recently echoed Szepesvári’s sentiments: “Reinforcement learning is the new wave of AI. More and more, we’re seeing companies exploring applications of reinforcement learning for process control and autonomous decision making. We’ve only just begun to scratch the surface of the value that can be created, and business leaders are starting to take notice.”

The popularity of the field is plain to those paying attention. Recently, Amii Chief Scientific Advisor Richard S. Sutton (also an Amii Fellow and Distinguished Research Scientist with DeepMind), released the second edition of his foundational textbook Reinforcement Learning: An Introduction. Szepesvári has also seen success with his course and online book Bandit Algorithms alongside Tor Lattimore.

Additionally, last year Amii and the University of Alberta launched the Reinforcement Learning Specialization on Coursera, taught by Martha White and Adam White, who are both Amii Fellows and Canada CIFAR AI Chairs in addition to being lab mates of Sutton and Szepesvári in UAlberta’s RLAI Lab.

The RL Theory Seminars, which are planned to run every Tuesday at 5:00 PM UTC, help researchers, practitioners and enthusiasts keep up with the pace of progress the field has seen over the past years. Each 50-minute talk (plus questions) will focus on the latest advances in theoretical reinforcement learning, taught by some of the world’s leading experts in the field. 

At a time when in-person interactions are limited, the seminars provide a platform for researchers to get together and discuss the latest work in RL theory. Organizers aim to provide a balanced view of contemporary RL theory and invite speakers covering a broad range of topics. Seminars begin on April 28 and are currently scheduled into mid-June.

Check out the first seminar, featuring Chi Jin of Princeton University, on April 28 at 5:00 PM UTC.