AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Stimulus domain transfer in recurrent models for large scale cortical population prediction on video

Fabian Sinz, Alexander S Ecker, Paul Fahey, Edgar Walker, Erick Cobos, Emmanouil Froudarakis, Dimitri Yatsenko, Zachary Pitkow, Jacob Reimer, Andreas Tolias

NeurIPS 2018 · Architectures and efficiency › RNNs and normalization · approach: Brain-inspired network comparison · 44 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
add shifter and modulator components (ablation)Recurrent core/readout model · Mouse V1 natural movies, scan 17358-5-3percent of oracle: 35.7% -> 50.4%Improved
add shifter and modulator components (ablation)Recurrent core/readout model · Mouse V1 natural movies, scan 17797-8-5percent of oracle: 56.3% -> 68.3%Improved
add shifter and modulator components (ablation)Recurrent core/readout model · Mouse V1 natural movies, scan 18142-6-3percent of oracle: 44.5% -> 57.7%Improved
use recurrent CNN+GRU core instead of 3D-conv linear-nonlinear model of similar parameter countNeural response predictor with shifter and modulator · Mouse V1 natural moviessingle-trial correlationImproved
add shifter and modulator to improve orientation tuning match on natural-movie model (ablation)Natural-movie trained model · Orientation tuning difference vs real neuronsvariance of preferred orientation difference: Levene p < 0.03, p < 0.0015, p < 10^-11Improved
train core on natural movies instead of noise, test on natural movies (ablation)Recurrent model · Cross-domain test on moviestest correlationImproved
train core on natural movies instead of noise, test on noise (ablation)Recurrent model · Cross-domain test on noisetest correlationNo clear effect
train readout on both domains instead of one (ablation)Natural-movie core · Noise and movie test setstest correlationImproved

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