Paper resolving a 15-year-old AdaBoost question, aided by LLMs, accepted to NeurIPS 2026
chetanp · x · 2026-11-19
Erik Ywang of Benchmark has a paper accepted to NeurIPS 2026 in Sydney: the work resolves an open question about AdaBoost's dynamics that stood for nearly 15 years, highlighting how LLMs can massively lower the barriers to frontier mathematical research. Benchmark is also hosting a lunch in Sydney on Dec 9 with frontier AI researchers and engineers.
More from Research
- RL never uses uniform sampling: bias sources and the RFCL fix — KyleMorgenstein · 2026-11-19
- Removing human video drops robot OOD success from 60% to 7%: generalization comes from human video — gui_penedo · 2026-11-19
- Grady Booch names frontier models' shared pattern: the 'Fallible Chaperone' — Grady_Booch · 2026-11-19
- Meta FAIR Montreal opens 2027 AIM residency applications for PhD students — _dsevero · 2026-11-19
- Provably distortion-free watermarking for any generative model via black-box access — serrjoa · 2026-11-19
- SolidWorks-Bench debuts: no AI model completes a single machine; Fable 5.2 leads at 57% — jxnlco · 2026-11-19