Yarin Gal at ECCV workshop: open problems in agent uncertainty quantification, from hallucinations to scientific discovery
abursuc · x · 2026-09-08
Oxford researcher Yarin Gal gave a talk on Open Problems in Agent Uncertainty Quantification at an ECCV 2026 workshop, spanning hallucinations, computer-use agents, and scientific discovery.
Core observation: discovery is a cycle — scientific advice rarely progresses linearly, and uncertainty drives experimentation. The poster notes it doubles as a good life/career observation.
Related event: Yarin Gal on Agent Uncertainty Quantification at ECCV(2 posts)→
More from Research
- Physicist: Navier-Stokes is physically meaningless at scales with possible singularities — MilesCranmer · 2026-09-08
- Denny Zhou: Least-to-Most Prompting Was an Early Agent Harness, Hitting 99% on SCAN — denny_zhou · 2026-09-08
- ECCV 2026 Workshop on Privacy-Preserving Visual Localization Features Brachmann, Pollefeys Talks — ducha_aiki · 2026-09-08
- MBZUAI launches Abu Dhabi deep phenotyping study for disease prediction — thamar_solorio · 2026-09-08
- Boundary-aware self-distillation precisely tunes LLM safety refusal boundaries — MultiverseComputingCAI · 2026-09-08
- Both Nvidia and Google run ARC synthetic data generators — GregKamradt · 2026-09-08