The year 2019 has seen major advances in self-supervised representation learning, with many new methods reaching high performances on standard benchmarks. Using better losses and augmentation methods, this trend will surely continue to slowly advance the field. However, it is also apparent that there are still major unresolved challenges and it is not clear what the next step-change is going to be. In this half-day workshop we want to highlight and provide a forum to discuss potential research direction seeds from radically new self-supervision tasks to downstream self-supervised learning and semi-supervised learning approaches.

As the methods are maturing, the field is now at the point where we have to start discussing how we can make optimal use of self-supervised representations in applications, as well as what are the remaining obstacles and possible approaches to tackle them. The workshop aims to give space to ask and discuss fundamental, longer-term questions with researchers that are leading this area. Key questions we aim to tackle include:

  • What can be learned by the current generation of self-supervised techniques? And what does, instead, still require manual supervision?
  • How can we make optimal use of self-supervised learning?
  • Is combining tasks the new way forward?
  • Are images enough? Video and multi-modal data as self-supervision is becoming popular.
  • Why do contrastive losses work well and do they scale?
  • How do we evaluate the quality of the learned representations?
  • How can we move to meaningful down-stream tasks that benefit from feature learning?
  • Why do methods such as clustering or contrastive losses work better than those that focus on interpretable, image-specific tasks as for example colorization or jigsaw puzzles?

Please note that this workshop will be a series of invited talks and does not have a call for paper submissions.


Tentative Schedule

All invited talks will be streamed and are followed by live QA with the speaker.

The talks are divided into two sessions of 2.5h to better accomodate the timesones of speakers and the audience. The first session starts on 09:00 and the second session 10h later on 19:00.

The times in the table below are automatically converted to the time zone of your browser.

09:00-09:05 Welcome
09:05-09:40 Invited Talk 1 Aäron van den Oord
09:40-10:15 Invited Talk 2 Paolo Favaro
10:15-10:50 Invited Talk 3 Carl Doersch
10:50-11:25 Invited Talk 4 Andrew Zisserman
19:00-19:05 Welcome
19:05-19:40 Invited Talk 5 Ishan Misra
19:40-20:15 Invited Talk 6 Stella Yu
20:15-20:50 Invited Talk 7 Alexei (Alyosha) Efros
20:50-21:25 Invited Talk 8 Deepak Pathak