mlsanityData to 7 Oct 2026

DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learning

arXiv 2006.08381 ↗First posted 15 Jun 2020cs.AI cs.LGcited by 159

Not on the map. Why?

Cited by 23 in LLM reasoning, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Expert problem-solving is driven by powerful languages for thinking about problems and their solutions. Acquiring expertise means learning these languages -- systems of concepts, alongside the skills to use them. We present DreamCoder, a system that learns to solve problems by writing programs. It builds expertise by creating programming languages for expressing domain concepts, together with neural networks to guide the search for programs within these languages. A ``wake-sleep'' learning algorithm alternately extends the language with new symbolic abstractions and trains the neural network on imagined and replayed problems. DreamCoder solves both classic inductive programming tasks and creative tasks such as drawing pictures and building scenes. It rediscovers the basics of modern functional programming, vector algebra and classical physics, including Newton's and Coulomb's laws. Concepts are built compositionally from those learned earlier, yielding multi-layered symbolic representations that are interpretable and transferrable to new tasks, while still growing scalably and flexibly with experience.

Key facts. Cited by 159 papers in the map (data through 7 Oct 2026). First posted on arXiv 15 Jun 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 24; 2022: 18; 2023: 30; 2024: 30; 2025: 29; 2026: 28.

159 papers.

Sort:

By area:

  1. 2021Program Synthesis with Large Language Modelscited by 3,509
  2. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,081
  3. 2024Agent Workflow Memorycited by 342
  4. 2023Language to Rewards for Robotic Skill Synthesiscited by 307
  5. 2023Hypothesis Search: Inductive Reasoning with Language Modelscited by 119
  6. 2023From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thoughtcited by 118
  7. 2023LEGO-Prover: Neural Theorem Proving with Growing Librariescited by 108
  8. 2021Abstraction and Analogy-Making in Artificial Intelligencecited by 103
  9. 2025Inducing Programmatic Skills for Agentic Taskscited by 103
  10. 2024WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environmentcited by 99
  11. 2023The ConceptARC Benchmark: Evaluating Understanding and Generalization in the ARC Domaincited by 97
  12. 2024Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systemscited by 94
38 more
  1. 2023Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinementcited by 89
  2. 2021Improving Coherence and Consistency in Neural Sequence Models with Dual-System, Neuro-Symbolic Reasoningcited by 85
  3. 2023LLMs and the Abstraction and Reasoning Corpus: Successes, Failures, and the Importance of Object-based Representationscited by 81
  4. 2024Formal Mathematical Reasoning: A New Frontier in AIcited by 80
  5. 2024Symbolic Regression with a Learned Concept Librarycited by 76
  6. 2024TroVE: Inducing Verifiable and Efficient Toolboxes for Solving Programmatic Taskscited by 70
  7. 2024Hypothesis Generation with Large Language Modelscited by 68
  8. 2021Learning to Synthesize Programs as Interpretable and Generalizable Policiescited by 59
  9. 2022Top-Down Synthesis for Library Learningcited by 57
  10. 2021Incorporating Domain Knowledge into Deep Neural Networkscited by 56
  11. 2022Structured, flexible, and robust: benchmarking and improving large language models towards more human-like behavior in out-of-distribution reasoning taskscited by 56
  12. 2024What Are Tools Anyway? A Survey from the Language Model Perspectivecited by 53
  13. 2023Large Language Models Are Not Strong Abstract Reasonerscited by 52
  14. 2021Engineering Sketch Generation for Computer-Aided Designcited by 49
  15. 2022Predicate Invention for Bilevel Planningcited by 49
  16. 2023From task structures to world models: What do LLMs know?cited by 46
  17. 2021Leveraging Language to Learn Program Abstractions and Search Heuristicscited by 44
  18. 2023Break It Down: Evidence for Structural Compositionality in Neural Networkscited by 44
  19. 2023Provable Compositional Generalization for Object-Centric Learningcited by 41
  20. 2023Learning adaptive planning representations with natural language guidancecited by 36
  21. 2021Communicating Natural Programs to Humans and Machinescited by 35
  22. 2021Fast and flexible: Human program induction in abstract reasoning taskscited by 33
  23. 2024Data-driven Discovery with Large Generative Modelscited by 33
  24. 2022Graphs, Constraints, and Search for the Abstraction and Reasoning Corpuscited by 32
  25. 2024Can Large Language Models Understand Symbolic Graphics Programs?cited by 32
  26. 2023Outline, Then Details: Syntactically Guided Coarse-To-Fine Code Generationcited by 27
  27. 2021Noether Networks: Meta-Learning Useful Conserved Quantitiescited by 26
  28. 2021Symbolic Behaviour in Artificial Intelligencecited by 24
  29. 2022Are Deep Neural Networks SMARTer than Second Graders?cited by 24
  30. 2021Neural-guided, Bidirectional Program Search for Abstraction and Reasoningcited by 23
  31. 2024CodeIt: Self-Improving Language Models with Prioritized Hindsight Replaycited by 23
  32. 2026DeltaEvolve: Accelerating Scientific Discovery through Momentum-Driven Evolutioncited by 23
  33. 2023Hierarchical Programmatic Reinforcement Learning via Learning to Compose Programscited by 20
  34. 2021ReaSCAN: Compositional Reasoning in Language Groundingcited by 19
  35. 2022Abstraction for Deep Reinforcement Learningcited by 19
  36. 2023GENOME: GenerativE Neuro-symbOlic visual reasoning by growing and reusing ModulEscited by 18
  37. 2024Pix2Code: Learning to Compose Neural Visual Concepts as Programscited by 18
  38. 2025Proving the Coding Interview: A Benchmark for Formally Verified Code Generationcited by 18

50 most cited of 159 All 159 (JSON)