The 35th conference on Neural Information Processing Systems (NeurIPS) runs Monday, December 6 through Tuesday, December 14, 2021. NeurIPS is a multi-track interdisciplinary annual meeting featuring the most significant achievements in the field of machine learning. The meeting begins with tutorials, followed by the main program with invited talks, oral and poster sessions, competition tracks, as well as demos, socials, and other events, and concludes with technical workshops. In 2021 NeurIPS is a virtual-only conference.
IARAI is excited to present its Traffic4cast competition as a part of the NeurIPS conference for the third consecutive year. This year’s contest titled “Temporal and Spatial Few-Shot Transfer Learning in Traffic Map Movie Forecasting” builds on the insights from the previous competitions and provides a larger urban traffic dataset in a more efficient encoding. Traffic4cast 2021 challenges machine learners to build traffic models robust to domain shifts in space and time and comprises the core and extended competitions.
A summary of Traffic4cast 2021 will be presented at the NeurIPS Competition Track on Thursday, December 9 at 19:00 CET (10:00 am PST). During this session, the winners of the core and extended competitions will give a short talk describing their solutions.
An in-depth discussion of the competition results will be presented at the Traffic4cast 2021 Special Session on Friday, December 10. The session consists of two parts starting at 10:00 CET (1:00 am PST) and at 13:30 CET (4:30 am PST). This event is free and open to everyone (NeurIPS 2021 registration is not required). The Special Session will be opened by an introduction to the Traffic4cast challenge by IARAI founding director Sepp Hochreiter. The first and second parts of the session will begin with a keynote lecture by Johannes Brandstetter from Johannes Kepler University, Austria, and Amsterdam Machine Learning Lab, Netherlands, and Monika Sester from Leibniz University Hannover, Germany, respectively. Then the winners of the core and extended challenges, the winner of the special prize, and the participants selected by the Traffic4cast Scientific Committee will describe their solutions in detail.