Research map / Learning theory and algorithms / Prediction and decision losses
Behavior Learning (BL)
Zhenyao Ma, Yue Liang, Dongxu Li
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 |
|---|---|---|---|
| use BL instead of E-MLP at depth 1 | BL · AG News | accuracy (%): 86.05 -> 89.29 | Improved |
| use BL instead of E-MLP at depth 2 | BL · AG News | accuracy (%): 85.11 -> 87.72 | Improved |
| use BL instead of E-MLP at depth 3 | BL · Yelp | accuracy (%): 89.59 -> 90.99 | Improved |
| use BL instead of E-MLP at depth 2 | BL · Fashion-MNIST OOD | AUROC (%): 84.61 -> 89.87 | Improved |
| use BL instead of E-MLP at depth 1 | BL · Yelp OOD | AUROC (%): 79.35 -> 84.31 | Improved |
| use BL instead of E-MLP at depth 1 | BL · AG News OOD | AUROC (%): 87.95 -> 83.15 | Hurt |
| use BL instead of E-MLP at depth 3 | BL · MNIST | accuracy (%): 98.14 -> 97.93 | No clear effect |
| use BL instead of E-MLP | BL · MNIST OOD | AUROC (%): 87.76 -> 92.92 | Mixed |
| lower temperature tau or raise penalty scale lambda | BL · energy-conservation constraint test | constraint violation |T|: within 10^-2 error at lambda = 25, tau = 0.01 | Improved |
| use BL(Shallow) instead of MLP | BL · 10 tabular datasets | F1-Macro rank | Improved |
Closest papers by meaning
- Inverse Decision Modeling: Learning Interpretable Representations of Behavior (ICML 2021 · similarity 0.85)
- Behavior Alignment via Reward Function Optimization (NeurIPS 2023 · similarity 0.84)
- SMILE: Bridging Continuous Optimization and Discrete Symbolic Recovery (NeurIPS 2026 · similarity 0.84)
- SYMBOL: Generating Flexible Black-Box Optimizers through Symbolic Equation Learning (ICLR 2024 · similarity 0.84)
- Solver-Informed RL: Grounding Large Language Models for Authentic Optimization Modeling (NeurIPS 2025 · similarity 0.83)
- Discriminative Policy Optimization for Token-Level Reward Models (ICML 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.
