Google DeepMind Paper Finds Two Numerical Errors in Standard Softmax and Cross-Entropy
burkov · x · 2026-10-25
- Burkov shares a Google DeepMind article identifying two numerical errors in standard softmax and cross-entropy calculations that can distort neural network training even when the underlying formulas are mathematically correct.
- One error removes the gradient for highly confident correct predictions while leaving small gradients for other classes; the other produces inaccurate, sometimes nonzero gradients.
- He also corrected the link: an internal URL was posted by mistake, and no ChapterPal account is needed to read the paper.
More from Research
- Paper: "bottleneck pressure" metric shows where AI-infeasible work gains value as AI speeds up the rest — soumitrashukla9 · 2026-10-28
- New IEEE IT paper: geometric Jensen-Shannon divergence between Gaussian measures on Hilbert space — FrnkNlsn · 2026-10-28
- One-page Q-learning notes with math, code and a playground for RL beginners — East-Muffin-6472 · 2026-10-28
- CPR-Dataset: retrievers fail at cross-domain discovery, training gives +70% on BRIGHT — coallaoh · 2026-10-28
- Hugging Face robot duck walks in simulated snow with 290,000 physics-simulated flakes — lukas_m_ziegler · 2026-10-28
- Mercor, Stanford and NBER launch TaskNet, an open-source task catalog updating O*NET for the AI era — soumitrashukla9 · 2026-10-28