mlsanityData to 7 Oct 2026

Predicate Invention for Bilevel Planning

arXiv 2203.09634 ↗First posted 17 Mar 2022cs.AI cs.LG cs.ROcited by 49

Not on the map. Why?

Abstract (arXiv)

Efficient planning in continuous state and action spaces is fundamentally hard, even when the transition model is deterministic and known. One way to alleviate this challenge is to perform bilevel planning with abstractions, where a high-level search for abstract plans is used to guide planning in the original transition space. Previous work has shown that when state abstractions in the form of symbolic predicates are hand-designed, operators and samplers for bilevel planning can be learned from demonstrations. In this work, we propose an algorithm for learning predicates from demonstrations, eliminating the need for manually specified state abstractions. Our key idea is to learn predicates by optimizing a surrogate objective that is tractable but faithful to our real efficient-planning objective. We use this surrogate objective in a hill-climbing search over predicate sets drawn from a grammar. Experimentally, we show across four robotic planning environments that our learned abstractions are able to quickly solve held-out tasks, outperforming six baselines. Code: https://tinyurl.com/predicators-release

Key facts. Cited by 49 papers in the map (data through 7 Oct 2026). Builds on 19 of its 60 references that are in the map. First posted on arXiv 17 Mar 2022.

Builds on

19 of 60 references.

References · 19

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2018Relational inductive biases, deep learning, and graph networkscited by 1,341
  3. 2015Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)cited by 1,061
  4. 2020Learning Invariant Representations for Reinforcement Learning without Reconstructioncited by 424
  5. 2020Integrated Task and Motion Planningcited by 281
  6. 2020DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learningcited by 159
  7. 2020Online Bayesian Goal Inference for Boundedly-Rational Planning Agentscited by 67
  8. 2017Near Optimal Behavior via Approximate State Abstractioncited by 63
  9. 2019Online Replanning in Belief Space for Partially Observable Task and Motion Problemscited by 60
  10. 2017Classical Planning in Deep Latent Space: Bridging the Subsymbolic-Symbolic Boundarycited by 58
  11. 2021Learning Symbolic Operators for Task and Motion Planningcited by 57
  12. 2020Learning compositional models of robot skills for task and motion planningcited by 47
7 more
  1. 2021Learning Neuro-Symbolic Relational Transition Models for Bilevel Planningcited by 43
  2. 2019Learning First-Order Symbolic Representations for Planning from the Structure of the State Spacecited by 38
  3. 2021Discovering State and Action Abstractions for Generalized Task and Motion Planningcited by 17
  4. 2019Generalized Lazy Search for Robot Motion Planning: Interleaving Search and Edge Evaluation via Event-based Togglescited by 16
  5. 2018Learning Quickly to Plan Quickly Using Modular Meta-Learningcited by 15
  6. 2021Extended Tree Search for Robot Task and Motion Planningcited by 14
  7. 2020Learning Neural-Symbolic Descriptive Planning Models via Cube-Space Priors: The Voyage Home (to STRIPS)cited by 7

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 6; 2024: 14; 2025: 11; 2026: 15.

49 papers.

Sort:

By area:

  1. 2023Generalized Planning in PDDL Domains with Pretrained Large Language Modelscited by 139
  2. 2024LLM3:Large Language Model-based Task and Motion Planning with Motion Failure Reasoningcited by 63
  3. 2024VisualPredicator: Learning Abstract World Models with Neuro-Symbolic Predicates for Robot Planningcited by 37
  4. 2022LEAGUE: Guided Skill Learning and Abstraction for Long-Horizon Manipulationcited by 32
  5. 2024Practice Makes Perfect: Planning to Learn Skill Parameter Policiescited by 22
  6. 2024A Survey of Optimization-based Task and Motion Planning: From Classical To Learning Approachescited by 19
  7. 2022Learning Efficient Abstract Planning Models that Choose What to Predictcited by 17
  8. 2023Embodied Lifelong Learning for Task and Motion Planningcited by 17
  9. 2024Partially Observable Task and Motion Planning with Uncertainty and Risk Awarenesscited by 17
  10. 2024Leveraging Environment Interaction for Automated PDDL Translation and Planning with Large Language Modelscited by 17
  11. 2025Learning Compositional Behaviors from Demonstration and Languagecited by 15
  12. 2025Synthesizing world models for bilevel planningcited by 13
37 more
  1. 2024From Pixels to Predicates: Learning Symbolic World Models via Pretrained Vision-Language Modelscited by 11
  2. 2024Open-World Task and Motion Planning via Vision-Language Model Generated Constraintscited by 9
  3. 2023Embodied Active Learning of Relational State Abstractions for Bilevel Planningcited by 8
  4. 2025ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planningcited by 8
  5. 2022Meta-Learning Parameterized Skillscited by 7
  6. 2025Language Models For Generalised PDDL Planning: Synthesising Sound and Programmatic Policiescited by 7
  7. 2025Hybrid Diffusion for Simultaneous Symbolic and Continuous Planningcited by 6
  8. 2023Generalisation Through Negation and Predicate Inventioncited by 5
  9. 2024Graph Learning for Numeric Planningcited by 5
  10. 2025Neuro-Symbolic Imitation Learning: Discovering Symbolic Abstractions for Skill Learningcited by 4
  11. 2023Discovering Predictive Relational Object Symbols with Symbolic Attentive Layerscited by 3
  12. 2024RoboMP2: A Robotic Multimodal Perception-Planning Framework with Multimodal Large Language Modelscited by 3
  13. 2025SLAP: Shortcut Learning for Abstract Planningcited by 3
  14. 2026Kintsugi: Learning Policies by Repairing Executable Knowledge Basescited by 3
  15. 2024CLIMB: Language-Guided Continual Learning for Task Planning with Iterative Model Buildingcited by 2
  16. 2025Few-Shot Neuro-Symbolic Imitation Learning for Long-Horizon Planning and Actingcited by 2
  17. 2023Learning logic programs by discovering higher-order abstractionscited by 1
  18. 2024Learning to Bridge the Gap: Efficient Novelty Recovery with Planning and Reinforcement Learningcited by 1
  19. 2024Neuro-Symbolic Skill Discovery for Conditional Multi-Level Planningcited by 1
  20. 2025Learning Hierarchical Domain Models Through Environment-Grounded Interactioncited by 1
  21. 2025Coloring Between the Lines: Personalization in the Null Space of Planning Constraintscited by 1
  22. 2024Sketch-Plan-Generalize: Learning and Planning with Neuro-Symbolic Programmatic Representations for Inductive Spatial Conceptscited by 0
  23. 2025Rational Inverse Reasoning: Few-Shot Imitation by Inferring Intent through Planningcited by 0
  24. 2026Build on Priors: Vision--Language--Guided Neuro-Symbolic Imitation Learning for Data-Efficient Real-World Robot Manipulationcited by 0
  25. 2026GRAIL: Autonomous Concept Grounding for Neuro-Symbolic Reinforcement Learningcited by 0
  26. 2026Neurosymbolic Imitation Learning with Human Guidance: A Privileged Information Approachcited by 0
  27. 2026Learning Bilevel Policies over Symbolic World Models for Long-Horizon Planningcited by 0
  28. 2026Neuro-Symbolic Learning for Long-Horizon Task Planning Under Complex Logical Constraintscited by 0
  29. 2026SCOPE: Evolving Symbolic World for Planning in Open-Ended Environmentscited by 0
  30. 2026Agent vs. Parametric World Models: Hybrid Planning for Reliable Language Agentscited by 0
  31. 2026NeSy-CSA: A Neuro-Symbolic Framework for Open-Ended Critical Scenario Attributioncited by 0
  32. 2026PDDL-ART: Autonomous Symbolic Abstraction From Demonstration For Long-Horizon Robotic Manipulation Using Vision-Language Modelscited by 0
  33. 2026Macro-Operator Generation and Predicate Selection for TAMP Operator Learningcited by 0
  34. 2026What Is Worth Representing? Representational Empowerment for Continual Model Constructioncited by 0
  35. 2026Managing Action Preconditions in Neuro-Symbolic RL: Three Placement Strategies for Embodied Agentscited by 0
  36. 2026StageGuard: Learning Stage Transitions for Long-Horizon Robot Tasks via Agentic Distillationcited by 0
  37. 2026Going Beyond State-Reaching: Learning Abstractions for Intrinsically Motivated Option Discoverycited by 0