srush: numerical formal verification of core ML code is easier and cheaper than expected
NLP researcher Sasha Rush (srush) argues that numerically verifying most core ML code via formal methods is both useful and feasible, and far cheaper than expected—cheaper than continual learning—citing a Google paper as motivation.
2026-11-14 ~ 2026-11-14 · 2 related posts
- Sasha Rush: numerically verifying core ML code is easier and cheaper than you think — srush_nlp · 2026-11-14
- srush: numerically verifying core ML code is doable — and cheaper than continual learning — srush_nlp · 2026-11-14