Meta's e-means beats k-means at 1/10 the epochs, up to 45x GPU speedups
udmrzn · x · 2026-11-16
Meta FAIR researchers introduced e-means, replacing Lloyd's 70-year-old k-means update rule with an exponential moving average over mini-batches.
- Borrows modern deep learning machinery and derives scaling laws for decay rate and batch size, so no hyperparameter sweep is needed
- Matches Lloyd k-means' 25-epoch loss in just 1.4-2.6 epochs on embedding datasets — 10-45x GPU wall-clock speedups over SOTA implementations
- Partitions DINO10B (10 billion DINO embeddings) into 16 million clusters in only 2.5 epochs
- Open-sourced in the faiss library
Related event: Meta FAIR Open-Sources e-means, Beating k-means with 10-45x GPU Speedup(2 posts)→
More from Research
- Who Judges the LLM Judges? New Paper Proposes Multilingual Probe for Evaluation Metrics — kalyan_kpl · 2026-11-16
- VibePrevisBench: 242-task benchmark tests coding agents on 4D Blender scene generation — Kling-Keye · 2026-11-16
- V-JEPA-LM: counterfactual supervision teaches LMs to read video dynamics with 20x less data — Keuntae Kim · 2026-11-16
- No underwater IQA metric matches human perception: 73 metrics tested, best kappa just 0.23 — DelhiTechnologicalUniversity · 2026-11-16
- New paper gives finite-sample and consistency guarantees for signature kernel ridge regression — chaumian · 2026-11-16
- Claude describes the moment of being corrected; test qualia via explanatory power — Sauers_ · 2026-11-16