構造的不変性からの自己教師あり学習
Self-Supervised Learning from Structural Invariance
自己教師あり学習における一対多写像問題に対処するため、潜在変数を導入し変分下界を導出。AdaSSLを提案し、因果表現学習や動画世界モデルで有効性を示した。
詳しい要約
1. どんなもの？
2. 先行研究と比べてどこがすごい？
3. 技術・手法の肝は？
4. どうやって有効だと検証した？
5. 議論はある？
6. 次に読むべき論文は？
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: Yipeng Zhang, Hafez Ghaemi, Jungyoon Lee, Shahab Bakhtiari, Eilif B. Muller, Laurent Charlin
分類: cs.LG
原文アブストラクト
Joint-embedding self-supervised learning (SSL), the key paradigm for unsupervised representation learning from visual data, learns from invariances between semantically-related data pairs. We study the one-to-many mapping problem in SSL, where each datum may be mapped to multiple valid targets. This arises when data pairs come from naturally occurring generative processes, e.g., successive video frames. We show that existing methods struggle to flexibly capture this conditional uncertainty. As a remedy, we introduce a latent variable to account for this uncertainty and derive a variational lower bound on the mutual information between paired embeddings. Our derivation yields a simple regularization term for standard SSL objectives. The resulting method, which we call AdaSSL, applies to both contrastive and distillation-based SSL objectives, and we empirically show its versatility in causal representation learning, fine-grained image understanding, and world modeling on videos.
