AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Cross-Subject Modeling for Widefield Calcium Imaging via Atlas-Aligned Spatiotemporal Tokenization

Mohammad Hosseini, Eray Erturk, Saba Hashemi, Maryam M. Shanechi

ICML 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Zero-shot robustness and transfer · 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
add global spatial embeddings (ablation)WiCAT (linear probing) · Musall zero-shotR2: 0.1985 -> 0.3274Improved
add global spatial embeddings (ablation)WiCAT (linear probing) · Kondo zero-shotR2: 0.1316 -> 0.1840Improved
use shared parameters instead of session-specific parameters (ablation)WiCAT (linear probing) · Musall zero-shotR2: 0.1025 -> 0.3274Improved
use shared parameters instead of session-specific parameters (ablation)WiCAT (linear probing) · Kondo zero-shotR2: 0.0301 -> 0.1840Improved
add self-supervised multi-subject pretraining (ablation)WiCAT · Musall zero-shotR2: 0.0936 -> 0.3274Improved
add self-supervised multi-subject pretraining (ablation)WiCAT · Kondo zero-shotR2: 0.0985 -> 0.1840Improved
use masked autoencoding instead of JEPA pretraining (ablation)WiCAT · Musall zero-shotR2: 0.2604 -> 0.3274Improved
use masked autoencoding instead of JEPA pretraining (ablation)WiCAT · Musall finetune setR2: 0.4980 -> 0.5124Improved
pretrain on multiple subjects instead of single session (ablation)WiCAT · Musall zero-shotR2: 0.1249 -> 0.3274Improved
replace linear decoder with Kalman smoother on frozen embeddingsWiCAT · Kondo zero-shotR2: 0.1840 -> 0.2869Improved
replace linear decoder with Kalman smoother on frozen embeddingsWiCAT · Musall zero-shotR2: 0.3274 -> 0.4336Improved
full finetuning of backbone instead of linear probingWiCAT · Kondo zero-shotR2: 0.1840 -> 0.2398Improved

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