Fixing intra-batch similarity nearly restores iid performance in streaming learning
y_m_asano · x · 2026-10-09
Building on Joao Carreira's streaming-training direction, the authors show that of sliding-window batching's two pathologies, only high intra-batch similarity is detrimental; fixing it nearly restores iid-sampling performance. They scale to a 100-hour concatenated Walking Tour stream, and it also works on urban dashcams, KrishnaCAM, and HD-EPIC.
Related event: StreamMAE Makes Self-Supervised Learning Work on Continuous Video Streams(4 posts)→
More from Research
- Formalized classification of finite simple groups may be near, says Irving — geoffreyirving · 2026-10-10
- Success-Guided Sampling: zero real-world data for dexterous robot hands — milesmacklin · 2026-10-10
- Google DeepMind launches AI co-clinician initiative for 'triadic care' in medicine — alan_karthi · 2026-10-10
- Anthropic's Geoffrey Irving: is the formalized classification of finite simple groups coming next week? — geoffreyirving · 2026-10-10
- Cell-rejuvenating therapy shows early vision improvement in 2 of 3 glaucoma patients — kimmonismus · 2026-10-10
- Terence Tao: The Era of "Big M" Mathematics and "Big P" Physics Has Arrived — CatAstro_Piyush · 2026-10-10