Timezone: »

Self-Supervised Learning -- Theory and Practice
Pengtao Xie · Shanghang Zhang · Pulkit Agrawal · Ishan Misra · Cynthia Rudin · Abdelrahman Mohamed · Wenzhen Yuan · Barret Zoph · Laurens van der Maaten · Xingyi Yang · Eric Xing

Sat Dec 12 08:50 AM -- 06:40 PM (PST) @
Event URL: https://sslneuips20.github.io/ »

Self-supervised learning (SSL) is an unsupervised approach for representation learning without relying on human-provided labels. It creates auxiliary tasks on unlabeled input data and learns representations by solving these tasks. SSL has demonstrated great success on images (e.g., MoCo, PIRL, SimCLR) and texts (e.g., BERT) and has shown promising results in other data modalities, including graphs, time-series, audio, etc. On a wide variety of tasks, SSL without using human-provided labels achieves performance that is close to fully supervised approaches.

The existing SSL research mostly focuses on improving the empirical performance without a theoretical foundation. While the proposed SSL approaches are empirically effective, theoretically why they perform well is not clear. For example, why certain auxiliary tasks in SSL perform better than others? How many unlabeled data examples are needed by SSL to learn a good representation? How is the performance of SSL affected by neural architectures?

In this workshop, we aim to bridge this gap between theory and practice. We bring together SSL-interested researchers from various domains to discuss the theoretical foundations of empirically well-performing SSL approaches and how the theoretical insights can further improve SSL’s empirical performance. Different from previous SSL-related workshops which focus on empirical effectiveness of SSL approaches without considering their theoretical foundations, our workshop focuses on establishing the theoretical foundation of SSL and providing theoretical insights for developing new SSL approaches.
We invite submissions of both theoretical works and empirical works, and the intersection of the two. The topics include but are not limited to:
Theoretical foundations of SSL
Sample complexity of SSL methods
Theory-driven design of auxiliary tasks in SSL
Comparative analysis of different auxiliary tasks
Comparative analysis of SSL and supervised approaches
Information theory and SSL
SSL for computer vision, natural language processing, robotics, speech processing, time-series analysis, graph analytics, etc.
SSL for healthcare, social media, neuroscience, biology, social science, etc.
Cognitive foundations of SSL

In addition to invited talks by leading researchers from diverse backgrounds including CV, NLP, robotics, theoretical ML, etc., the workshop will feature poster sessions and panel discussion to share perspectives on establishing foundational understanding of existing SSL approaches and theoretically-principled ways of developing new SSL methods. We accept submissions of short papers (up to 4 pages excluding references in NeurIPS format), which will be peer-reviewed by at least two reviewers. The accepted papers are allowed to be submitted to other conference venues.

Author Information

Pengtao Xie (Carnegie Mellon University)
Shanghang Zhang (UC Berkeley)
Pulkit Agrawal (MIT)
Ishan Misra (Facebook AI Research)
Cynthia Rudin (Duke)
Abdelrahman Mohamed (Facebook AI Research (FAIR))
Wenzhen Yuan
Barret Zoph (Google Research, Brain Team)
Laurens van der Maaten (Facebook)
Xingyi Yang (University of California San Diego)
Eric Xing (Petuum Inc. / Carnegie Mellon University)

More from the Same Authors