AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

Torus Graphs for Large Scale Neural Phase Analysis

Jack Goffinet, Casey Hanks, David Carlson

ICML 2026 · Probabilistic and scientific modeling › Dynamical systems · approach: Functional connectivity modeling

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 exact closed-form score matching with stochastic score matching (ablation)Torus Graph · Synthetic TG data, high dimensionparameter recovery R^2 / feasible dimension: exact fails at d~100 (memory); stochastic works at d in thousandsImproved
use stochastic instead of exact score matching in low dimensions (ablation)Torus Graph · Synthetic TG data, low dimensionparameter recovery R^2No clear effect
use stochastic score matching inside the TG-HMM (ablation)TG-HMM · Synthetic HMM state recoverystate accuracy: accurate beyond d~100 where exact is memory-limitedImproved
fit AR-TG with score matching instead of maximum likelihood (ablation)AR-TG · Bivariate causal direction, syntheticdirection accuracyImproved
use AR-TG instead of baseline TE estimators (Kraskov, KDE, KL, Granger)Transfer entropy estimation · Bivariate synthetic causal directiondirection accuracyImproved
use AR-TG instead of multivariate Granger causalityCausal discovery · Synthetic multivariate, varying dimension and sparsityedge recovery AUC / runtime: Granger 30-hour timeout past d=64; AR-TG under one hourImproved
model conditional independence with TG-HMM instead of pairwise PLVSpindle analysis · Mouse LFP spindles (1,334 events)interpretability of coupling changesImproved
include phase-sum terms in TGTorus Graph · Mouse LFP Wake/NREMparameter magnitudeNo 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