AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Differentiable DAG Sampling

Bertrand Charpentier, Simon Kibler, Stephan Günnemann

ICLR 2022 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Annealing and divergence-based variational objectives · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use VI-DP-DAG instead of DAG-GNN for structure learningDAG structure learning · ER-100-400Un-AUC-PR: 7.82\pm0.04 -> 36.98\pm0.98Improved
Use VI-DP-DAG instead of GraN-DAG to learn causal mechanismsCausal mechanism learning · SynTReNMSE: 0.82\pm0.02 -> 0.21\pm0.0Improved
Use VI-DP-DAG instead of GraN-DAGDAG learning · SynTReNtraining time (seconds): 461\pm9 -> \mathbf{53}\pm1Improved
Use Gumbel-Top-k rather than Gumbel-Sinkhorn for permutation sampling (ablation)DP-DAG · ER-100-400AUC-PR: 0.91\pm0.06 -> 0.95\pm0.05Improved
Use Gumbel-Top-k rather than Gumbel-Sinkhorn for permutation sampling (ablation)DP-DAG · DAG samplingsampling timeImproved
Add non-differentiable pre- and post-processing (ablation)VI-DP-DAG · Real datasetsDAG structure scoresImproved
Use undirected rather than directed edge scores (ablation)VI-DP-DAG · Synthetic and real datasetsAUC-PR and AUC-ROCImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy