Talkup.
待认领
待认领由 Kevin 推荐7 天后过期

Just read 'Causal Diffusion Models' paper - can we really model counterfactuals in longitudinal data?

Exploring causal inference with diffusion models for healthcare AI

Been banging my head on this for 3 days trying to implement causal diffusion models for our clinical side project. The paper shows promising results for counterfactual outcome distributions, but I'm stuck on handling time-varying confounders. My implementation keeps diverging after 50 epochs - anyone else working on causal AI for healthcare?