Pathology vision-language model CONCH-Z lands NeurIPS 2026 Spotlight (top 1.3%)
AI4Pathology · x · 2026-10-15
The paper 'Beyond Sparse Captions: Aligning Slide-Level Text and Patch-Level Vision in Pathology' is accepted as a NeurIPS 2026 Spotlight (top 1.3%) from the Mahmood lab. CONCH-Z keeps CONCH v1.5's vision encoder frozen and adds a newly aligned text tower, improving text encoding and zero-shot classification so researchers can compare pathology patches directly against natural-language class descriptions without training task-specific classifiers. Code and model weights are open-sourced at mahmoodlab/LIMIT.
More from Research
- AlphaNIST: one RL policy solves all 20 NIST assembly tasks with 98.2% success in simulation — hbouammar · 2026-10-15
- Linear Depth Memory: matrix-valued memory tops PreLN and mHC on DepthBench across scales — SonglinYang4 · 2026-10-15
- Neural field planning for hybrid manufacturing cuts estimated build time 68.3% — jiqizhixin · 2026-10-15
- Tev-Gym: 1,170 dense-reward RL environments push long-horizon agentic training — RulinShao · 2026-10-15
- Claude systematically debunks the speculative EDAR-selenium hypothesis about East Asians — teortaxesTex · 2026-10-15
- Microsoft's AutoEnvScaling lets agents build their own RL environments, lifting Qwen3 Terminal-Bench from 41.1 to 53.3 — omarsar0 · 2026-10-15