Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Modeling Task Effects on Meaning Representation in the Brain via Zero-Shot MEG Prediction
Mariya Toneva, Otilia Stretcu, Barnabas Poczos, Leila Wehbe, Tom M. Mitchell
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Add task (question) semantics to MEG prediction (additive H3 or precomputed attention H4.1) versus stimulus-only H1 (ablation) | Ridge regression on MTurk features · MEG 2v2 accuracy, 475-550 ms | 2v2 accuracy | Improved |
| Add task semantics to stimulus-only model (ablation) | Ridge regression on MTurk features · MEG 2v2 accuracy, early windows (before 450 ms) | 2v2 accuracy | No clear effect |
| Learned attention (H4.2) instead of precomputed attention (H4.1) (ablation) | Interactive task-stimulus model · MEG 2v2 accuracy | 2v2 accuracy | No clear effect |
| Interactive attention (H4.1) instead of additive (H3) task-stimulus combination (ablation) | Ridge regression on MTurk features · MEG 2v2 accuracy, averaged over sensors | 2v2 accuracy | No clear effect |
| Use BERT word and question features instead of MTurk representations (ablation) | All hypotheses · MEG 2v2 accuracy | 2v2 accuracy | Hurt |
| Replace stimulus representations with random vectors (ablation) | Ridge regression · MEG 2v2 accuracy | 2v2 accuracy: chance (0.5) | Hurt |
| Increase amount of training data (ablation) | H3 and H4.2 · MEG 2v2 accuracy | 2v2 accuracy | Improved |
Closest papers by meaning
- Inducing brain-relevant bias in natural language processing models (NeurIPS 2019 · similarity 0.82)
- Can fMRI reveal the representation of syntactic structure in the brain? (NeurIPS 2021 · similarity 0.81)
- Divergences between Language Models and Human Brains (NeurIPS 2024 · similarity 0.80)
- fMRI predictors based on language models of increasing complexity recover brain left lateralization (NeurIPS 2024 · similarity 0.80)
- Bridging the Semantic Latent Space between Brain and Machine: Similarity Is All You Need (AAAI 2024 · similarity 0.79)
- Far from the Shallow: Brain-Predictive Reasoning Embedding through Residual Disentanglement (NeurIPS 2025 · similarity 0.79)
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.
