New XAI study links latent-space causality to real flow structures via SHAP
ricardovinuesa · x · 2026-11-04
A new XAI study connects disentangled autoencoder latent variables to physical flow features:
- SHAP attributions identify which flow structures each latent variable controls
- This bridges causal analysis performed in latent space with actual flow structures decoded back to the field
- Pipeline: explainability (encoder) → causality in latent space → decoder mapping back to flow
Paper details at the link in the original post.
More from Research
- Ai2 releases StateBench: Gated DeltaNet tracks state over sequences 10× longer than training — allen_ai · 2026-11-04
- AI research meme: no one can beat your method if you never release the dataset — miniapeur · 2026-11-04
- Native image reading beats OCR for deep research agents, cutting tool calls — RulinShao · 2026-11-04
- CMU's IdeaTaste benchmark: LLM research taste reaches only ~26% of human level — CMU-SCS · 2026-11-04
- JFPuget Team Takes 3rd on Public Leaderboard of Both Kaggle ARC-AGI Competitions — JFPuget · 2026-11-04
- Stanford Data-Centric AI Workshop 2026 kicks off with full-day live talks — jure · 2026-11-04