mlsanityData to 14 Oct 2026

Goal-Aware Prediction: Learning to Model What Matters

arXiv 2007.07170 ↗First posted 14 Jul 2020cs.LG cs.AI cs.ROcited by 45

Not on the map. Why?

Abstract (arXiv)

Learned dynamics models combined with both planning and policy learning algorithms have shown promise in enabling artificial agents to learn to perform many diverse tasks with limited supervision. However, one of the fundamental challenges in using a learned forward dynamics model is the mismatch between the objective of the learned model (future state reconstruction), and that of the downstream planner or policy (completing a specified task). This issue is exacerbated by vision-based control tasks in diverse real-world environments, where the complexity of the real world dwarfs model capacity. In this paper, we propose to direct prediction towards task relevant information, enabling the model to be aware of the current task and encouraging it to only model relevant quantities of the state space, resulting in a learning objective that more closely matches the downstream task. Further, we do so in an entirely self-supervised manner, without the need for a reward function or image labels. We find that our method more effectively models the relevant parts of the scene conditioned on the goal, and as a result outperforms standard task-agnostic dynamics models and model-free reinforcement learning.

Key facts. Cited by 45 papers in the map (data through 14 Oct 2026). First posted on arXiv 14 Jul 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 14; 2022: 9; 2023: 7; 2024: 7; 2025: 2; 2026: 6.

45 papers.

Sort:

By area:

  1. 2022R3M: A Universal Visual Representation for Robot Manipulationcited by 694
  2. 2022Goal-Conditioned Reinforcement Learning: Problems and Solutionscited by 158
  3. 2021Physion: Evaluating Physical Prediction from Vision in Humans and Machinescited by 153
  4. 2025Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Modelscited by 89
  5. 2023DREAMWALKER: Mental Planning for Continuous Vision-Language Navigationcited by 85
  6. 2021Greedy Hierarchical Variational Autoencoders for Large-Scale Video Predictioncited by 75
  7. 2021Language Grounding with 3D Objectscited by 43
  8. 2023Contrastive Difference Predictive Codingcited by 36
  9. 2022Generalizing Goal-Conditioned Reinforcement Learning with Variational Causal Reasoningcited by 34
  10. 2022Value Gradient weighted Model-Based Reinforcement Learningcited by 32
  11. 2022Deciding What to Model: Value-Equivalent Sampling for Reinforcement Learningcited by 19
  12. 2021VisuoSpatial Foresight for Physical Sequential Fabric Manipulationcited by 18
33 more
  1. 2022Control-Aware Prediction Objectives for Autonomous Drivingcited by 18
  2. 2023Plan To Predict: Learning an Uncertainty-Foreseeing Model for Model-Based Reinforcement Learningcited by 17
  3. 2023Learning to Discern: Imitating Heterogeneous Human Demonstrations with Preference and Representation Learningcited by 15
  4. 2024Solving Minimum-Cost Reach Avoid using Reinforcement Learningcited by 14
  5. 2025Efficient Robotic Policy Learning via Latent Space Backward Planningcited by 14
  6. 2021MapGo: Model-Assisted Policy Optimization for Goal-Oriented Taskscited by 13
  7. 2021Learning Domain Invariant Representations in Goal-conditioned Block MDPscited by 12
  8. 2021LS3: Latent Space Safe Sets for Long-Horizon Visuomotor Control of Sparse Reward Iterative Taskscited by 10
  9. 2021DisCo RL: Distribution-Conditioned Reinforcement Learning for General-Purpose Policiescited by 9
  10. 2024Pre-trained Visual Dynamics Representations for Efficient Policy Learningcited by 8
  11. 2026Beyond Next-Observation Prediction: Agent-Authored World Modeling for Sequential Decision Makingcited by 8
  12. 2021Goal-Aware Cross-Entropy for Multi-Target Reinforcement Learningcited by 7
  13. 2022On Rate-Distortion Theory in Capacity-Limited Cognition & Reinforcement Learningcited by 6
  14. 2024Learning World Models for Unconstrained Goal Navigationcited by 6
  15. 2024Policy-shaped prediction: avoiding distractions in model-based reinforcement learningcited by 6
  16. 2024Tell Me What's Next: Textual Foresight for Generic UI Representationscited by 5
  17. 2024Make the Pertinent Salient: Task-Relevant Reconstruction for Visual Control with Distractionscited by 5
  18. 2021Learning Intuitive Physics with Multimodal Generative Modelscited by 4
  19. 2023Bayesian Reinforcement Learning with Limited Cognitive Loadcited by 4
  20. 2024BECAUSE: Bilinear Causal Representation for Generalizable Offline Model-based Reinforcement Learningcited by 3
  21. 2026ProWorld: Progress-Aware Hyperbolic World Models for Long-Horizon Visual Goal Reachingcited by 3
  22. 2021Discriminator Augmented Model-Based Reinforcement Learningcited by 2
  23. 2023Minimal Value-Equivalent Partial Models for Scalable and Robust Planning in Lifelong Reinforcement Learningcited by 2
  24. 2021Model-Advantage and Value-Aware Models for Model-Based Reinforcement Learning: Bridging the Gap in Theory and Practicecited by 1
  25. 2021Comparing Reconstruction- and Contrastive-based Models for Visual Task Planningcited by 1
  26. 2026Data-Efficient Hierarchical Goal-Conditioned Reinforcement Learning via Normalizing Flowscited by 1
  27. 2021ED2: Environment Dynamics Decomposition World Models for Continuous Controlcited by 0
  28. 2022Task-Agnostic Learning to Accomplish New Taskscited by 0
  29. 2022Performance-Driven Controller Tuning via Derivative-Free Reinforcement Learningcited by 0
  30. 2023Imitator Learning: Achieve Out-of-the-Box Imitation Ability in Variable Environmentscited by 0
  31. 2026Improving Offline Goal-Conditioned Reinforcement Learning via Selective Reward Stimulationcited by 0
  32. 2026What Must a World Model Distinguish for Planning?cited by 0
  33. 2026Learning Macroscopic Dynamics without Reconstructing Microscopic Statescited by 0