Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Meta-Learning of Structured Task Distributions in Humans and Machines
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 |
|---|---|---|---|
| Evaluate humans on compositional rather than null boards | Humans · Compositional vs null task distributions | blue tiles revealed | Helped |
| Evaluate the A2C LSTM agent on null rather than compositional boards | A2C LSTM agent · Null vs compositional task distributions | blue tiles revealed | Helped |
| Compare human performance with the A2C LSTM agent | A2C LSTM agent · Compositional task distribution | blue tiles revealed | Hurt |
| Compare human performance with the A2C LSTM agent | A2C LSTM agent · Null task distribution | blue tiles revealed | Hurt |
| Replace the first fully connected layer with a convolutional layer (ablation) | A2C LSTM agent · Null task distribution | blue tiles revealed | Helped |
| Replace the first fully connected layer with a convolutional layer (ablation) | A2C LSTM agent · Compositional task distribution | blue tiles revealed | Hurt |
| Replace the first fully connected layer with a convolutional layer (ablation) | A2C LSTM agent · Compositional and null task distributions | blue tiles revealed | Mixed |
| Use the nearest-neighbor heuristic instead of human behavior | Humans · Compositional and null task distributions | blue tiles revealed | Hurt |
| Use the nearest-neighbor heuristic instead of the convolutional agent | Convolutional agent · Compositional and null task distributions | blue tiles revealed | Hurt |
| Split compositional trials into the first and last half | Humans · Compositional task distribution | success rate | Helped |
| Split null-distribution trials into the first and last half | A2C LSTM agent · Null task distribution | success rate | Helped |
Closest papers by meaning
- MetaCARD: Meta-Reinforcement Learning with Task Uncertainty Feedback via Decoupled Context-Aware Reward and Dynamics Components (AAAI 2024 · similarity 0.89)
- AMAGO-2: Breaking the Multi-Task Barrier in Meta-Reinforcement Learning with Transformers (NeurIPS 2024 · similarity 0.89)
- Human-like Category Learning by Injecting Ecological Priors from Large Language Models into Neural Networks (ICML 2024 · similarity 0.89)
- Decoupling Meta-Reinforcement Learning with Gaussian Task Contexts and Skills (AAAI 2024 · similarity 0.89)
- Meta-Reinforcement Learning with Adaptation from Human Feedback via Preference-Order-Preserving Task Embedding (ICML 2025 · similarity 0.88)
- Entropy Regularized Task Representation Learning for Offline Meta-Reinforcement Learning (AAAI 2025 · similarity 0.88)
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.