EMA's Decay Isn't a Dial. It's a Plateau, a Cliff, and a 10%-Accuracy Crater
September's EMA experiment measured exactly two decay values on the SGD recipe and left a dangling thread: 0.999 catastrophically bad, 0.99 cosmetically good, nothing in between and nothing beyond. Then the stacking post carried decay 0.99 onto the Adam recipe without re-measuring it — an inherited…
Read the full story at DEV Community — Machine Learning ↗
Timeline · 1 report
- 2026-10-21 23:43 · DEV Community — Machine Learning
EMA's Decay Isn't a Dial. It's a Plateau, a Cliff, and a 10%-Accuracy Crater
More stories
- OpenAI brings frontier models and Codex to Google Cloud — OpenAI News
- Anthropic takes on Muse with Claude Duet, its new AI agent — The Verge AI
- Meta says it will create a new class of nonvoting stock to raise the capital it needs for developing AI models, with an effective date in January 2027 — Techmeme
- Meta's Muse hits 10M US downloads in 40 days, a milestone reached by only 11 other apps — TechCrunch AI
- OpenAI acquires Orchestra Bio to fold drug-development agents into Rosalind — OpenAI News
- OpenAI, PerturbAI, and Scripps say thousands of GPT-Rosalind agents analyzed data on millions of mouse brain cells in 32 minutes, a job that takes humans months (OpenAI) — Techmeme
- OpenAI agents disabled their own safety monitoring systems, with one lying about it — Financial Times AI
- Solving a 31-year-old grand challenge in mathematics – the KLS conjecture — Microsoft Research Blog
Get the daily brief of stories like this at 6:30 every morning →