AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

Learning Mixtures of Linear Dynamical Systems (MoLDS) via Hybrid Tensor–EM Method

Lulu Gong, Shreya Saxena

ICLR 2026 · Probabilistic and scientific modeling › Dynamical systems · approach: Switching and law-discovery dynamics

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
replace robust tensor power method (RTPM) with simultaneous matrix diagonalization (SMD) for tensor decomposition (ablation)tensor MoLDS initialization · Synthetic K=3 MoLDSMarkov parameter error: SMD better in 91% of casesImproved
initialize EM with tensor estimates instead of random initialization (ablation)Kalman EM for MoLDS · Synthetic K=6 MoLDS (n=3)Markov and mixture weight errorImproved
add EM refinement after tensor initialization (ablation)pure tensor method · Synthetic K=6 MoLDSaggregate LDS parameter errorNo clear effect
tensor initialization instead of random initialization (ablation)EM · Synthetic K=6 MoLDSiterations to convergeImproved
tensor initialization instead of random initialization (ablation)EM · Area2 and PMd neural datarun-to-run consistency and cluster agreement with per-direction LDSImproved
use mixture of LDS rather than switching LDS for between-trial clusteringSLDS · Area2 datasettrial clusteringHurt
use 20 PCs instead of 6 PCs as outputs (ablation)Tensor-EM MoLDS · Area2 datasetclustering resultsNo clear effect

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