New paper: Two routes to scalable credit assignment without weight symmetry, code released
aran_nayebi · x · 2026-09-16
Aran Nayebi, Daniel Kunin, Surya Ganguli and collaborators released "Two Routes to Scalable Credit Assignment without Weight Symmetry", exploring two paths to scalable credit assignment that don't rely on weight symmetry. Paper link, code and weights are publicly available, with the author pointing to earlier bioplausible-learning-on-ImageNet work for context.
More from Research
- Six years on, scvi-tools still widely used — outlasting foundation models and coding agents — anshulkundaje · 2026-09-16
- Lightning Weave composes reasoning capabilities via on-policy distillation — Yecheng Wu · 2026-09-16
- Danish administrative data finds no effect of AI on earnings or wages — paulnovosad · 2026-09-16
- Zero Agent Trained on Synthetic Data From a Simulated World of Talking Characters — generativist · 2026-09-16
- NeurIPS paper: value-based RL agents can fail to converge to any optimal policy in Newcomblike environments — jessi_cata · 2026-09-16
- DAI-S2S-ST: 153K Human Preference Ratings Rank Voice Models on How They Feel to Talk To — rdesh26 · 2026-09-16