AI Paper Atlas

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

A Normative and Biologically Plausible Algorithm for Independent Component Analysis

Yanis Bahroun, Dmitri Chklovskii, Anirvan M. Sengupta

NeurIPS 2021 · Learning theory and algorithms › Prediction and decision losses · approach: Theory of neural network sensitivity · 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 SM-ICA instead of pre-whitening-dependent algorithms on colored dataSM-ICA · Synthetic scenario 2 (colored sub-Gaussian)source recovery MSEImproved
Use SM-ICA on a mix of sub- and super-Gaussian sourcesSM-ICA · Synthetic scenario 3 (white, mixed kurtosis)source recovery MSEImproved
Use SM-ICA on colored data with mixed kurtosisSM-ICA · Synthetic scenario 4source recovery MSEImproved
Compare SM-ICA with Nonlinear Oja and Amari on white sub-Gaussian dataSM-ICA · Synthetic scenario 1convergence speedNo clear effect
Use EASI baseline on mixed-kurtosis sourcesEASI · Synthetic scenarios 3 and 4source recoveryHurt
Use online instead of offline SM-ICA (ablation)SM-ICA · Speech separationresidualHurt
Use online instead of offline SM-ICA (ablation)SM-ICA · Image separationsource recoveryNo clear effect
Use fixed learning rates instead of time-varying (ablation)SM-ICA · All tasksseparation qualityNo clear effect
Apply SM-ICA to speech and image mixtures with no preprocessing changeSM-ICA · Speech (TSP) and natural imagessource recoveryImproved

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