AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

Do Not Marginalize Mechanisms, Rather Consolidate!

Moritz Willig, Matej Zečević, Devendra Singh Dhami, Kristian Kersting

NeurIPS 2023 · Probabilistic and scientific modeling › Causal discovery · approach: Counterfactual identification and bounds · 0 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
Consolidate instead of marginalize the intermediate variables of the tool-wear SCMTool-wear time-series SCM · Milling machine exampleability to evaluate interventionsImproved
Consolidate instead of marginalize the game-agent SCM, keeping only distance and planning_sequence variablesPlatformer agent SCM · Agent policy exampleinterpretability of the agent policyImproved
Remove child-less variables not in the consolidation setConsolidated SCM · Any SCMmodel complexityImproved
Consolidate chains by function composition under interventionsConsolidated SCM · Chainsequation simplificationMixed
Consolidate the parent of a forkConsolidated SCM · Forksequation sizeHurt
Consolidate both parents of a collider with shared parentsConsolidated SCM · Collidersequation simplificationImproved
Consolidate an already minimal SCMConsolidated SCM · Minimal systemsequation simplificationNo clear effect
Consolidate time steps of the tool-wear sharpness equationTool-wear SCM · Sharpness S_tclosed-form equation: (1 - 0.3 * 0.5)^t = 0.85^tImproved

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