Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Learning Awareness Models
Brandon Amos, Laurent Dinh, Serkan Cabi, Thomas Rothörl, Sergio Gómez Colmenarejo, Alistair Muldal, Tom Erez, Yuval Tassa, Nando de Freitas, Misha Denil
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 model-prediction entropy to actively collect training data rather than rely on passive grasp-and-release data collection | PreCo dynamics model for the simulated MPL hand · Simulated MPL hand, maximum fingertip pressure control | reward | Improved |
| Add temporally correlated action noise to planner-selected data-collection actions (ablation) | Planning-based data collection policies · Simulated MPL hand, maximum fingertip pressure control | reward | Improved |
| Use active entropy-based data collection instead of passive data collection | PreCo dynamics model with shape diagnostic · Simulated MPL hand, target-shape prediction when not in contact | classification loss | Improved |
| Train a forward model on proprioceptive and tactile observations, then read out object properties from its hidden states | PreCo dynamics model · Simulated MPL hand, target-shape prediction over an episode | shape prediction | Improved |
| Use PreCo hidden states rather than direct sensor readings as features for the diagnostic | PreCo dynamics model and diagnostic · Shadow Hand, grasped-object orientation prediction | orientation prediction error | Improved |
| Minimize predicted observation entropy rather than maximize it | PreCo model used for MPC · Simulated MPL hand, interaction with the target object | contact avoidance | Improved |
Closest papers by meaning
- Aligning Perception, Reasoning, Modeling and Interaction: A Survey on Physical AI (TPAMI 2026 · similarity 0.87)
- RoboScape: Physics-informed Embodied World Model (NeurIPS 2025 · similarity 0.85)
- Scaling Proprioceptive-Visual Learning with Heterogeneous Pre-trained Transformers (NeurIPS 2024 · similarity 0.85)
- SOLD: Slot Object-Centric Latent Dynamics Models for Relational Manipulation Learning from Pixels (ICML 2025 · similarity 0.84)
- Enhancing Tactile-based Reinforcement Learning for Robotic Control (NeurIPS 2025 · similarity 0.84)
- Neural Motion Simulator Pushing the Limit of World Models in Reinforcement Learning (CVPR 2025 · similarity 0.84)
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.
