AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Learning Mixtures of Markov Chains and MDPs

Chinmaya Kausik, Kevin Tan, Ambuj Tewari

ICML 2023 · Learning theory and algorithms › Prediction and decision losses · approach: Determinantal point process inference · 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
use the proposed subspace-estimation and clustering pipeline instead of randomly initialized soft EM8x8 gridworld mixture · 8x8 gridworld mixtureaverage accuracy: 73.2% -> 96.6%Improved
replace estimated projectors with the identity matrix (ablation)Proposed method · 8x8 gridworld mixtureclustering errorHurt
replace estimated projectors with a random projector (ablation)Proposed method · 8x8 gridworld mixtureclustering errorHurt
use estimated subspaces for trajectory clustering (ablation)Proposed method · 8x8 gridworld mixtureclustering errorImproved
increase trajectory length past the reported transition region (ablation)Proposed method · 8x8 gridworld mixtureclustering error: T_n=60 to T_n=70Improved
initialize soft EM with the proposed clustering and classification labels instead of random initialization (ablation)Soft EM · 8x8 gridworld mixturelabel accuracyImproved
initialize soft EM with models estimated from clustered trajectories instead of random initialization (ablation)Soft EM · 8x8 gridworld mixturelabel accuracy: T_n > 60Improved
initialize soft EM with clustering labels while introducing uncertainty in the responsibilities (ablation)Soft EM · 8x8 gridworld mixtureconvergenceImproved

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