Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Discovering and Achieving Goals via World Models
Russell Mendonca, Oleh Rybkin, Kostas Daniilidis, Danijar Hafner, Deepak Pathak
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 |
|---|---|---|---|
| remove separate explorer policy (ablation) | LEXA · RoboBins | success rate | Hurt |
| remove negative sampling for temporal distance (ablation) | LEXA · RoboBins | success rate | Hurt |
| train distance function on real data instead of imagination (ablation) | LEXA · RoboBins | success rate | Hurt |
| use temporal distance instead of cosine distance | LEXA · RoboKitchen | goal success rate (%): 6.02 -> 37.50 | Improved |
| use temporal distance instead of cosine distance | LEXA · RoboBins | goal success rate (%): 45.83 -> 69.44 | Improved |
| use cosine distance instead of temporal distance | LEXA · Walker | goal success rate (%): 36.72 -> 73.06 | Improved |
| use cosine distance instead of temporal distance | LEXA · Quadruped | goal success rate (%): 31.39 -> 56.11 | Improved |
| replace DDL with LEXA temporal | DDL · Walker | goal success rate (%): 40.00 -> 36.72 | Hurt |
| replace SkewFit with LEXA | Skew-Fit · SkewFit Pusher | goal distance: 4.9cm -> 2.3cm | Improved |
| replace DISCERN with LEXA | DISCERN · Pointmass | goal success (%): 49.6 -> 83.2 | Improved |
| replace Plan2Explore with LEXA | Plan2Explore · Reacher Hard | zero-shot return: 66 -> 937 | Improved |
| replace Plan2Explore with LEXA | Plan2Explore · Cartpole Balance | zero-shot return: 950 -> 886 | Hurt |
Closest papers by meaning
- GoalLadder: Incremental Goal Discovery with Vision-Language Models (NeurIPS 2025 · similarity 0.87)
- A Single Goal is All You Need: Skills and Exploration Emerge from Contrastive RL without Rewards, Demonstrations, or Subgoals (ICLR 2025 · similarity 0.86)
- Learning World Models for Unconstrained Goal Navigation (NeurIPS 2024 · similarity 0.85)
- Behavioral Exploration: Learning to Explore via In-Context Adaptation (ICML 2025 · similarity 0.85)
- RoboDreamer: Learning Compositional World Models for Robot Imagination (ICML 2024 · similarity 0.85)
- Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach (CVPR 2025 · similarity 0.85)
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.
