Research map / Label-efficient and robust learning / Recognition and long-tail
Beyond Hearing: Learning Task-agnostic ExG Representations from Earphones via Physiology-informed Tokenization
Hyungjun Yoon, Seungjoo Lee, Yu Wu, XiaoMeng Chen, Taiting Lu, Freddy Yifei Liu, Taeckyung Lee, Hyeongheon Cha, Haochen Zhao, Gaoteng Zhao, Dongyao Chen, Cecilia Mascolo and 2 more
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 free-living pre-training | PiMT · DailySense | avg F1: 81.9% to 87.6% | Improved |
| add free-living pre-training | PatchTST · DailySense | avg F1: 72.4% to 73.2% | Improved |
| add pre-training to Bidirectional-Mamba without PiMT | Bidirectional-Mamba · DailySense | avg F1: 0.779 to 0.813 | Improved |
| add pre-training to Bidirectional-Mamba without PiMT | Bidirectional-Mamba · DailySense | gaze angular error: 6.53 to 6.23 | Improved |
| add pre-training to Transformer without PiMT | Transformer · DailySense | avg F1: 0.724 to 0.732 | No clear effect |
| use Mamba backbone instead of Transformer (ablation) | PiMT · DailySense | F1: 6.4% improvement without pre-training and 8.5% with pre-training | Improved |
| use 12-band filter bank instead of 1/2/4 bands (ablation) | PiMT tokenization · DailySense | avg F1: average 4.6% F1-score gain | Improved |
| add autoencoding pre-training task (ablation) | PiMT · DailySense | avg F1: 0.873 -> 0.876 | Improved |
| add masked reconstruction pre-training task (ablation) | PiMT · DailySense | gaze angular error: 6.10 -> 6.00 | Improved |
| add amplitude reconstruction pre-training task (ablation) | PiMT · DailySense | gaze angular error: 6.19 -> 6.00 | Improved |
| add phase reconstruction pre-training task (ablation) | PiMT · DailySense | gaze angular error: 5.98 -> 6.00 | No clear effect |
| use PiMT tokenization instead of PatchTST | PiMT · DailySense (no pre-training) | avg F1: 0.724 -> 0.819 | Improved |
Closest papers by meaning
- S$^2$M-Former: Spiking Symmetric Mixing Branchformer for Brain Auditory Attention Detection (NeurIPS 2025 · similarity 0.85)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.84)
- CSBrain: A Cross-scale Spatiotemporal Brain Foundation Model for EEG Decoding (NeurIPS 2025 · similarity 0.84)
- What do EEG Foundation Models Capture from Human Brain Signals? (NeurIPS 2026 · similarity 0.84)
- FEEL: Quantifying Heterogeneity in Physiological Signals for Generalizable Emotion Recognition (NeurIPS 2025 · similarity 0.83)
- LUNA: Efficient and Topology-Agnostic Foundation Model for EEG Signal Analysis (NeurIPS 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.
