mlsanityData to 13 Oct 2026

Dreamer-CDP: Improving Reconstruction-free World Models Via Continuous Deterministic Representation Prediction

arXiv 2603.07083 ↗First posted 7 Mar 2026cs.LGcited by 2

On the map: World models → Latent dynamics models and model-based RL

also in 1 more
Abstract (arXiv)

Model-based reinforcement learning (MBRL) agents operating in high-dimensional observation spaces, such as Dreamer, rely on learning abstract representations for effective planning and control. Existing approaches typically employ reconstruction-based objectives in the observation space, which can render representations sensitive to task-irrelevant details. Recent alternatives trade reconstruction for auxiliary action prediction heads or view augmentation strategies, but perform worse in the Crafter environment than reconstruction-based methods. We close this gap between Dreamer and reconstruction-free models by introducing a JEPA-style predictor defined on continuous, deterministic representations. Our method matches Dreamer's performance on Crafter, demonstrating effective world model learning on this benchmark without reconstruction objectives.

Key facts. Cited by 2 papers in the map (data through 13 Oct 2026). Placed in Latent dynamics models and model-based RL, one of 11 directions in World models. Builds on 34 of its 38 references that are in the map. First posted on arXiv 7 Mar 2026.

Builds on

34 of 38 references.

Closest ancestors · 10

Same direction, newest first

  1. 2025Learning Transformer-based World Models with Contrastive Predictive Coding6 Mar
  2. 2024MuDreamer: Learning Predictive World Models without Reconstruction23 May
  3. 2024EfficientZero V2: Mastering Discrete and Continuous Control with Limited Data1 Mar
  4. 2023TD-MPC2: Scalable, Robust World Models for Continuous Control25 Oct
  5. 2023Mastering Diverse Domains through World Models10 Jan
  6. 2022DreamingV2: Reinforcement Learning with Discrete World Models without Reconstruction1 Mar
  7. 2021Mastering Atari Games with Limited Data30 Oct
  8. 2021DreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical Representations27 Oct
  9. 2019Dream to Control: Learning Behaviors by Latent Imagination3 Dec
  10. 2018World Models27 Mar

Other references · 24

  1. 2018Representation Learning with Contrastive Predictive Codingcited by 7,712
  2. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,730
  3. 2020Exploring Simple Siamese Representation Learningcited by 2,860
  4. 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,850
  5. 2019Mastering Atari, Go, Chess and Shogi by Planning with a Learned Modelcited by 1,334
  6. 2018Learning Latent Dynamics for Planning from Pixelscited by 1,241
  7. 2021VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learningcited by 1,033
  8. 2025V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningcited by 769
  9. 2024DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planningcited by 369
  10. 2020Data-Efficient Reinforcement Learning with Self-Predictive Representationscited by 311
  11. 2019DeepMDP: Learning Continuous Latent Space Models for Representation Learningcited by 205
  12. 2021Benchmarking the Spectrum of Agent Capabilitiescited by 194
12 more
  1. 2025Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Modelscited by 84
  2. 2024Learning and Leveraging World Models in Visual Representation Learningcited by 72
  3. 2022BYOL-Explore: Exploration by Bootstrapped Predictioncited by 71
  4. 2023TACO: Temporal Latent Action-Driven Contrastive Loss for Visual Reinforcement Learningcited by 56
  5. 2021Temporal Predictive Coding For Model-Based Planning In Latent Spacecited by 54
  6. 2024Bridging State and History Representations: Understanding Self-Predictive RLcited by 52
  7. 2022Understanding Self-Predictive Learning for Reinforcement Learningcited by 41
  8. 2023Contrastive Difference Predictive Codingcited by 35
  9. 2023Curious Replay for Model-based Adaptationcited by 16
  10. 2024When does Self-Prediction help? Understanding Auxiliary Tasks in Reinforcement Learningcited by 13
  11. 2024A Unifying Framework for Action-Conditional Self-Predictive Reinforcement Learningcited by 7
  12. 2024Focus On What Matters: Separated Models For Visual-Based RL Generalizationcited by 6

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Delta-JEPA: Learning Action-Sensitive World Models via Latent Difference Decodingcited by 15
  2. 2026Learning Commute-Time-Preserving World Models for Planningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2020Dreaming: Model-based Reinforcement Learning by Latent Imagination without Reconstruction
  2. 2025Zero-shot World Models via Search in Memory
  3. 2024Learning Latent Dynamic Robust Representations for World Models
  4. 2026Next Embedding Prediction Makes World Models Stronger
  5. 2026Reinformed Dreamer: An Asymmetric World Model Efficiently Trained through Latent Guidance
  6. 2026R2-Dreamer: Redundancy-Reduced World Models without Decoders or Augmentation
  7. 2020Mastering Atari with Discrete World Models
  8. 2026CausalDreamer: Learning Predictive World Models with Latent Disentanglement