mlsanityData to 14 Oct 2026

Hyperspherical Contrastive Representations Unlock Planning and Exploration

arXiv 2610.20409 ↗First posted 14 Oct 2026cs.LGcited by 0

On the map: Deep reinforcement learning methods → RL for robot locomotion and manipulation also fits: RL for robot navigation and UAV control

Abstract (arXiv)

We develop a self-supervised reinforcement learning algorithm that learns to build towers of up to eight cubes from scratch, without prior data, rewards, demonstrations, or pretrained models. Our method, Contrastive Planning via Spherical Linear Interpolation (C-Slerp), uses contrastive representations as reusable building blocks and repeatedly composes them to infer a sequence of intermediate waypoints. Under our hyperspherical parameterization, the learned representations allow intermediate waypoints to be inferred in closed form through spherical linear interpolation, without training a separate high-level policy or performing graph search. Importantly, the representation geometry also tells us how to adjust the number of waypoints based on the difficulty of each goal-reaching problem. Experiments show that our method succeeds on long-horizon navigation and robotic manipulation tasks, including settings where strong baselines fail to achieve a single success. Ablations further show that using a fixed number of waypoints substantially degrades performance, highlighting the importance of adapting the plan to task difficulty.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods. Builds on 27 of its 34 references that are in the map. First posted on arXiv 14 Oct 2026.

Builds on

27 of 34 references.

References · 27

  1. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
  2. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,405
  3. 2018Soft Actor-Critic Algorithms and Applicationscited by 1,240
  4. 2017Curiosity-driven Exploration by Self-supervised Predictioncited by 1,194
  5. 2017Hindsight Experience Replaycited by 1,136
  6. 2018Exploration by Random Network Distillationcited by 993
  7. 2019On Variational Bounds of Mutual Informationcited by 559
  8. 2018Data-Efficient Hierarchical Reinforcement Learningcited by 455
  9. 2017FeUdal Networks for Hierarchical Reinforcement Learningcited by 337
  10. 2022Contrastive Learning as Goal-Conditioned Reinforcement Learningcited by 253
  11. 2019Search on the Replay Buffer: Bridging Planning and Reinforcement Learningcited by 194
  12. 2019Planning with Goal-Conditioned Policiescited by 149
15 more
  1. 2023HIQL: Offline Goal-Conditioned RL with Latent States as Actionscited by 141
  2. 2021Goal-Conditioned Reinforcement Learning with Imagined Subgoalscited by 134
  3. 2020World Model as a Graph: Learning Latent Landmarks for Planningcited by 69
  4. 2019Why Does Hierarchy (Sometimes) Work So Well in Reinforcement Learning?cited by 68
  5. 2020Generating Adjacency-Constrained Subgoals in Hierarchical Reinforcement Learningcited by 58
  6. 20251000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilitiescited by 45
  7. 2024A Single Goal is All You Need: Skills and Exploration Emerge from Contrastive RL without Rewards, Demonstrations, or Subgoalscited by 23
  8. 2021C-Planning: An Automatic Curriculum for Learning Goal-Reaching Taskscited by 17
  9. 2024Inference via Interpolation: Contrastive Representations Provably Enable Planning and Inferencecited by 17
  10. 2025Horizon Generalization in Reinforcement Learningcited by 17
  11. 2025Flattening Hierarchies with Policy Bootstrappingcited by 11
  12. 2019Spherical Principal Component Analysiscited by 10
  13. 2025Contrastive Representations for Temporal Reasoningcited by 6
  14. 2025BuilderBench: The Building Blocks of Intelligent Agentscited by 6
  15. 2025Demystifying the Mechanisms Behind Emergent Exploration in Goal-conditioned RLcited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Outperforming 1000 Euclidean Layers With Hyperbolic Goal-Conditioned RL
  2. 2026Direction-Conditioned Policies for Online Goal-Conditioned Reinforcement Learning
  3. 2024Probabilistic World Modeling with Asymmetric Distance Measure
  4. 2026Three Steps at a Time: Learning Representations from Action Sequences in Contrastive RL
  5. 2024Contrastive Abstraction for Reinforcement Learning
  6. 2024Contrastive Diffuser: Planning Towards High Return States via Contrastive Learning
  7. 2025Equivariant Goal Conditioned Contrastive Reinforcement Learning
  8. 2026Learning Object Manipulation from Scratch via Contrastive Interaction