mlsanityData to 14 Oct 2026

DAGR: State-Conditioned Goal Representations via Difference-Aware Goal Cross-Attention

arXiv 2607.13731 ↗First posted 15 Jul 2026cs.LG stat.MLcited by 0

On the map: Deep reinforcement learning methods → Deep RL algorithms and representations

Abstract (arXiv)

Goal-conditioned reinforcement learning hinges on how the goal is encoded. Contrastive, metric, temporal-distance and information-theoretic encoders disagree on the objective. They agree on one thing. None of them sees the current state, so the embedding cannot mark which part of the goal still needs action, and the policy must recover that cue by inverting both encoders. We propose DAGR, which refines the static embedding of any late-fusion encoder into a state-conditioned one through multi-scale gated cross-attention. A gated residual holds the refinement near the base, and a difference-aware attention rule biases the scores by a per-token state-goal mismatch. A single condition decides what such a refinement can guarantee, namely whether the block returns its input at closed gates. We prove that the usual post-norm placement violates it, measure the consequence on frozen checkpoints, and recover part of the resulting loss by restoring the condition. On OGBench DAGR improves navigation and matches or trails the base elsewhere. Our ablations trace the gain to the gated residual rather than to the difference bias that names the method. Code is available at https://github.com/leixingxing1/DAGR

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods. Builds on 12 of its 64 references that are in the map. First posted on arXiv 15 Jul 2026.

Builds on

12 of 64 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025GCHR : Goal-Conditioned Hindsight Regularization for Sample-Efficient Reinforcement Learning8 Aug
  2. 2025Flattening Hierarchies with Policy Bootstrapping20 May

Other references · 10

  1. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  2. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  3. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  4. 2020Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problemscited by 1,920
  5. 2024Octo: An Open-Source Generalist Robot Policycited by 1,484
  6. 2022Goal-Conditioned Reinforcement Learning: Problems and Solutionscited by 158
  7. 2022How Far I'll Go: Offline Goal-Conditioned Reinforcement Learning via f-Advantage Regressioncited by 66
  8. 2026ACoT-VLA: Action Chain-of-Thought for Vision-Language-Action Modelscited by 33
  9. 2025Option-aware Temporally Abstracted Value for Offline Goal-Conditioned Reinforcement Learningcited by 23
  10. 2026NFTR: From Provable Mode-Averaging to Geodesic Subgoal Selection in Offline Goal-Conditioned RLcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Dual Goal Representations
  2. 2026MSPR: Multi-scale Predictive Representations for Goal-conditioned Reinforcement Learning
  3. 2026Dual Advantage Fields
  4. 2026Do Better Goal Representations Improve Goal-Conditioned Reinforcement Learning?
  5. 2023Goal-conditioned Offline Reinforcement Learning through State Space Partitioning
  6. 2026Functional Graphs for Predicting and Explaining Goal Failure in Sparse Goal-Conditioned RL
  7. 2026Direction-Conditioned Policies via Compositional Subgoal Scoring for Online Goal-Conditioned Reinforcement Learning
  8. 2026PRTS: A Primitive Reasoning and Tasking System via Contrastive Representations