Research map / Label-efficient and robust learning / Recognition and long-tail
GLOBEM Dataset: Multi-Year Datasets for Longitudinal Human Behavior Modeling Generalization
Xuhai Xu, Han Zhang, Yasaman S Sefidgar, Yiyi Ren, Xin Liu, Woosuk Seo, Jennifer Brown, Kevin Scott Kuehn, Mike A Merrill, Paula S Nurius, Shwetak Patel, Tim Althoff and 4 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 Reorder order-reconstruction self-supervised pretext task to 1D-CNN backbone | ERM-1dCNN · Leave-one-dataset-out depression detection | balanced accuracy: 0.510 -> 0.547 | Improved |
| add Reorder pretext task to 1D-CNN backbone | ERM-1dCNN · Single dataset past/future | balanced accuracy: 0.568 -> 0.626 | Improved |
| add Mixup augmentation to ERM | ERM-1dCNN · Leave-one-dataset-out | balanced accuracy: 0.510 -> 0.501 | No clear effect |
| use person as domain in DANN instead of dataset | DANN · Single dataset past/future | balanced accuracy: 0.526 -> 0.502 | Hurt |
| move from single-dataset to cross-dataset evaluation | Xu et al.-I · depression detection | balanced accuracy: 0.691 -> 0.502 | Hurt |
| move from single-dataset to cross-dataset evaluation | Reorder · depression detection | balanced accuracy: 0.626 -> 0.547 | Hurt |
| apply domain generalization algorithms (IRM, DANN, CSD, MLDG, MASF) over ERM | ERM-1dCNN · Leave-one-dataset-out | balanced accuracy: 0.510 vs 0.506-0.521 | No clear effect |
| normalize features per individual before domain classification | Random Forest · Name-the-dataset classification | accuracy | Mixed |
Closest papers by meaning
- Beyond Sensor Data: Foundation Models of Behavioral Data from Wearables Improve Health Predictions (ICML 2025 · similarity 0.87)
- FEEL: Quantifying Heterogeneity in Physiological Signals for Generalizable Emotion Recognition (NeurIPS 2025 · similarity 0.84)
- OpenMHC: Accelerating the Science of Wearable Foundation Models (NeurIPS 2026 · similarity 0.84)
- UniMTS: Unified Pre-training for Motion Time Series (NeurIPS 2024 · similarity 0.83)
- Using Adaptive Bandit Experiments to Increase and Investigate Engagement in Mental Health (AAAI 2024 · similarity 0.82)
- Towards Real-world Human Behavior Simulation: Benchmarking Large Language Models on Long-horizon, Cross-scenario, Heterogeneous Behavior Traces (NeurIPS 2026 · similarity 0.82)
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.
