Lab lands 9 NeurIPS 2026 papers, from multi-agent self-evolution to motion modeling
flosalim · x · 2026-09-25
Researcher flosalim announces 9 accepted NeurIPS 2026 papers from their group, promising a write-up for each. Main track: FOGO (forgetting-aware orthogonalization optimizer), GlucoFM (with the Google team), MAGE (multi-agent self-evolution with co-evolutionary knowledge graphs), selective-deployment residual correction for time series, AnyMo (geometry-aware in-the-wild human motion modeling), programmatic traffic-signal policies, and OOD detection in continual learning. Two evaluation-track papers cover retrieval-reranking root cause analysis and graded cognitive diversity in LLM persona simulation.
Related event: One team lands 9 papers at NeurIPS 2026(3 posts)→
More from Research
- Aesthetic valuation shaped by culturally-mediated tuning of perception, akin to language acquisition — abenitezburraco · 2026-09-25
- PosteriorBench accepted to NeurIPS: point reconstruction accuracy misleads on inverse solvers — AnimaAnandkumar · 2026-09-25
- Dietterich: pretraining's next-token task already forces models to abstract and generalize — tdietterich · 2026-09-25
- Neural Spectral Capacity: a training-free architecture score that beats #Params and prunes LLaMA-7B 5900x faster — CityU-HongKong · 2026-09-25
- HRI 2027 Adds Archival Industry White Paper Track for Real-World Robot Deployments — petitegeek · 2026-09-25
- Study: Reasoning hurts 15.7% of multimodal embeddings; training-free SURE router fixes it — _reachsumit · 2026-09-25