Research map / Large language models / Language, safety and interpretability
Learning to Predict Readability Using Eye-Movement Data From Natives and Learners
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 multi-task eye-movement prediction (GECO L1 native gaze) as auxiliary task | MLP readability model · OneStopEnglish I-E | accuracy: 59.30 -> 62.64 | Helped |
| add multi-task gaze prediction (GECO L2 learner gaze, 1st pass) | MLP readability model · Simple vs Normal Wikipedia | accuracy: 85.95 -> 86.62 | Helped |
| add multi-task gaze prediction (GECO L1 regression) | MLP readability model · OneStopEnglish A-E | accuracy: 67.53 -> 69.18 | Helped |
| use native (L1) instead of learner (L2) gaze as auxiliary task (ablation) | multi-task MLP · OneStopEnglish I-E | accuracy: 62.64 (L1 1st pass) vs 61.84 (L2 1st pass) | No clear effect |
| use GECO gaze instead of Dundee gaze as auxiliary data (ablation) | multi-task MLP · Simple vs Normal Wikipedia | accuracy: 86.45 -> 86.62 | Helped |
| multi-task gaze auxiliary task with only 100 training sentences | multi-task MLP · Simple vs Normal Wikipedia | accuracy: about 10% more accurate than single-task | Helped |
Closest papers by meaning
- CLIP-Gaze: Towards General Gaze Estimation via Visual-Linguistic Model (AAAI 2024 · similarity 0.84)
- Voila-A: Aligning Vision-Language Models with User's Gaze Attention (NeurIPS 2024 · similarity 0.84)
- MCIF: Multimodal Crosslingual Instruction-Following Benchmark from Scientific Talks (ICLR 2026 · similarity 0.84)
- Eyes on VLM: Benchmarking Gaze Following and Social Gaze Prediction in Vision Language Models (NeurIPS 2026 · similarity 0.84)
- Unveiling Visual Perception in Language Models: An Attention Head Analysis Approach (CVPR 2025 · similarity 0.83)
- EyEar: Learning Audio Synchronized Human Gaze Trajectory Based on Physics-Informed Dynamics (AAAI 2025 · similarity 0.83)
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.