mlsanityData to 14 Oct 2026

Learning What Matters: Adaptive Information-Theoretic Objectives for Robot Exploration

arXiv 2605.12084 ↗First posted 12 May 2026cs.RO cs.AI cs.ITcited by 0

Not on the map. Why?

Abstract (arXiv)

Designing learnable information-theoretic objectives for robot exploration remains challenging. Such objectives aim to guide exploration toward data that reduces uncertainty in model parameters, yet it is often unclear what information the collected data can actually reveal. Although reinforcement learning (RL) can optimize a given objective, constructing objectives that reflect parametric learnability is difficult in high-dimensional robotic systems. Many parameter directions are weakly observable or unidentifiable, and even when identifiable directions are selected, omitted directions can still influence exploration and distort information measures. To address this challenge, we propose Quasi-Optimal Experimental Design (Qfootnotesize OED), an adaptive information objective grounded in optimal experimental design. Qfootnotesize OED (i) performs eigenspace analysis of the Fisher information matrix to identify an observable subspace and select identifiable parameter directions, and (ii) modifies the exploration objective to emphasize these directions while suppressing nuisance effects from non-critical parameters. Under bounded nuisance influence and limited coupling between critical and nuisance directions, Qfootnotesize OED provides a constant-factor approximation to the ideal information objective that explores all parameters. We evaluate Qfootnotesize OED on simulated and real-world navigation and manipulation tasks, where identifiable-direction selection and nuisance suppression yield performance improvements of SI35.23percent and SI21.98percent, respectively. When integrated as an exploration objective in model-based policy optimization, Qfootnotesize OED further improves policy performance over established RL baselines.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 54 of its 75 references that are in the map. First posted on arXiv 12 May 2026.

Builds on

54 of 75 references.

References · 54

  1. 2017Attention Is All You Needcited by 60,361
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
  4. 2018Neural Ordinary Differential Equationscited by 3,779
  5. 2017Curiosity-driven Exploration by Self-supervised Predictioncited by 1,194
  6. 2018Exploration by Random Network Distillationcited by 993
  7. 2018Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Modelscited by 774
  8. 2019When to Trust Your Model: Model-Based Policy Optimizationcited by 773
  9. 2017A Distributional Perspective on Reinforcement Learningcited by 765
  10. 2016Unifying Count-Based Exploration and Intrinsic Motivationcited by 639
  11. 2023TD-MPC2: Scalable, Robust World Models for Continuous Controlcited by 512
  12. 2024One Step Diffusion via Shortcut Modelscited by 391
38 more
  1. 2020Planning to Explore via Self-Supervised World Modelscited by 368
  2. 2016#Exploration: A Study of Count-Based Exploration for Deep Reinforcement Learningcited by 309
  3. 2020First return, then explorecited by 266
  4. 2019Self-Supervised Exploration via Disagreementcited by 265
  5. 2017Deep Exploration via Randomized Value Functionscited by 156
  6. 2015Incentivizing Exploration In Reinforcement Learning With Deep Predictive Modelscited by 154
  7. 2023ANYmal Parkour: Learning Agile Navigation for Quadrupedal Robotscited by 152
  8. 2022DeXtreme: Transfer of Agile In-hand Manipulation from Simulation to Realitycited by 149
  9. 2024Reconciling Reality through Simulation: A Real-to-Sim-to-Real Approach for Robust Manipulationcited by 144
  10. 2021Discovering and Achieving Goals via World Modelscited by 131
  11. 2023Modern Bayesian Experimental Designcited by 125
  12. 2014Generalization and Exploration via Randomized Value Functionscited by 112
  13. 2024Humanoid Parkour Learningcited by 96
  14. 2018Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approachcited by 91
  15. 2021Understanding Domain Randomization for Sim-to-real Transfercited by 80
  16. 2018Information Directed Sampling and Bandits with Heteroscedastic Noisecited by 76
  17. 2018Efficient Exploration through Bayesian Deep Q-Networkscited by 67
  18. 2023FisherRF: Active View Selection and Uncertainty Quantification for Radiance Fields using Fisher Informationcited by 64
  19. 2019Distributional Reinforcement Learning for Efficient Explorationcited by 60
  20. 2023DTC: Deep Tracking Controlcited by 56
  21. 2018Information-Directed Exploration for Deep Reinforcement Learningcited by 52
  22. 2025Robotic World Model: A Neural Network Simulator for Robust Policy Optimization in Roboticscited by 51
  23. 2023Kronecker-Factored Approximate Curvature for Modern Neural Network Architecturescited by 49
  24. 2016Learning to Perform Physics Experiments via Deep Reinforcement Learningcited by 47
  25. 2024ASID: Active Exploration for System Identification in Robotic Manipulationcited by 28
  26. 2018DORA The Explorer: Directed Outreaching Reinforcement Action-Selectioncited by 27
  27. 2026mjlab: A Lightweight Framework for GPU-Accelerated Robot Learningcited by 26
  28. 2022Redeeming Intrinsic Rewards via Constrained Optimizationcited by 24
  29. 2021On structural and practical identifiabilitycited by 19
  30. 2025GauSS-MI: Gaussian Splatting Shannon Mutual Information for Active 3D Reconstructioncited by 17
  31. 2017Efficient exploration with Double Uncertain Value Networkscited by 16
  32. 2024Next Best Sense: Guiding Vision and Touch with FisherRF for 3D Gaussian Splattingcited by 10
  33. 2025Sampling-Based System Identification with Active Exploration for Legged Robot Sim2Real Learningcited by 8
  34. 2024You've Got to Feel It To Believe It: Multi-Modal Bayesian Inference for Semantic and Property Predictioncited by 7
  35. 2023Bayesian Active Learning in the Presence of Nuisance Parameterscited by 6
  36. 2024Adaptive Diffusion Terrain Generator for Autonomous Uneven Terrain Navigationcited by 6
  37. 2022Real-to-Sim: Predicting Residual Errors of Robotic Systems with Sparse Data using a Learning-based Unscented Kalman Filtercited by 5
  38. 2024Fast Extrinsic Calibration for Multiple Inertial Measurement Units in Visual-Inertial Systemcited by 4

50 most cited of 54 All 54 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025On Efficient Bayesian Exploration in Model-Based Reinforcement Learning
  2. 2024Geometric Active Exploration in Markov Decision Processes: the Benefit of Abstraction
  3. 2025An Information-Geometric Approach to Artificial Curiosity
  4. 2024Effective Exploration Based on the Structural Information Principles
  5. 2025A Novel Framework for Uncertainty-Driven Adaptive Exploration
  6. 2024Efficient Model-Based Reinforcement Learning Through Optimistic Thompson Sampling
  7. 2024Augmented Bayesian Policy Search
  8. 2024Bayesian Optimization Framework for Efficient Fleet Design in Autonomous Multi-Robot Exploration