mlsanityData to 7 Oct 2026

From Perception to Programs: Regularize, Overparameterize, and Amortize

arXiv 2206.05922 ↗First posted 13 Jun 2022cs.AIcited by 13

Not on the map. Why?

Abstract (arXiv)

Toward combining inductive reasoning with perception abilities, we develop techniques for neurosymbolic program synthesis where perceptual input is first parsed by neural nets into a low-dimensional interpretable representation, which is then processed by a synthesized program. We explore several techniques for relaxing the problem and jointly learning all modules end-to-end with gradient descent: multitask learning; amortized inference; overparameterization; and a differentiable strategy for penalizing lengthy programs. Collectedly this toolbox improves the stability of gradient-guided program search, and suggests ways of learning both how to perceive input as discrete abstractions, and how to symbolically process those abstractions as programs.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 30 of its 54 references that are in the map. First posted on arXiv 13 Jun 2022.

Builds on

30 of 54 references.

References · 30

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  4. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  5. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  6. 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,708
  7. 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,250
  8. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,037
  9. 2018Gradient Descent Finds Global Minima of Deep Neural Networkscited by 572
  10. 2015Importance Weighted Autoencoderscited by 454
  11. 2015Neural Module Networkscited by 375
  12. 2017Learning Explanatory Rules from Noisy Datacited by 210
18 more
  1. 2019BSP-Net: Generating Compact Meshes via Binary Space Partitioningcited by 205
  2. 2018Learning Equations for Extrapolation and Controlcited by 151
  3. 2017RobustFill: Neural Program Learning under Noisy I/Ocited by 124
  4. 2016Identity Matters in Deep Learningcited by 107
  5. 2017CSGNet: Neural Shape Parser for Constructive Solid Geometrycited by 106
  6. 2020UCSG-Net -- Unsupervised Discovering of Constructive Solid Geometry Treecited by 77
  7. 2017Learning to Infer Graphics Programs from Hand-Drawn Imagescited by 73
  8. 2018HOUDINI: Lifelong Learning as Program Synthesiscited by 38
  9. 2016Differentiable Programs with Neural Librariescited by 30
  10. 2020Learning Differentiable Programs with Admissible Neural Heuristicscited by 30
  11. 2021Techniques for Symbol Grounding with SATNetcited by 24
  12. 2019Synthesizing Datalog Programs Using Numerical Relaxationcited by 23
  13. 2020PLAD: Learning to Infer Shape Programs with Pseudo-Labels and Approximate Distributionscited by 19
  14. 2022CrossBeam: Learning to Search in Bottom-Up Program Synthesiscited by 18
  15. 2019Neural Shape Parsers for Constructive Solid Geometrycited by 14
  16. 2016TerpreT: A Probabilistic Programming Language for Program Inductioncited by 12
  17. 2021Web Question Answering with Neurosymbolic Program Synthesiscited by 11
  18. 2022Grammar-Based Grounded Lexicon Learningcited by 10

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 1; 2023: 2; 2024: 5; 2025: 3; 2026: 2.

13 papers.

Sort:

By area:

  1. 2024WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environmentcited by 99
  2. 2023Not All Neuro-Symbolic Concepts Are Created Equal: Analysis and Mitigation of Reasoning Shortcutscited by 63
  3. 2024Combining Induction and Transduction for Abstract Reasoningcited by 53
  4. 2024Pix2Code: Learning to Compose Neural Visual Concepts as Programscited by 18
  5. 2024Interpretable Concept-Based Memory Reasoningcited by 17
  6. 2023Neurosymbolic Grounding for Compositional World Modelscited by 14
  7. 2025Shortcuts and Identifiability in Concept-based Models from a Neuro-Symbolic Lenscited by 13
  8. 2025ARC-GEN: A Mimetic Procedural Benchmark Generator for the Abstraction and Reasoning Corpuscited by 8
  9. 2024Learning to Edit Visual Programs with Self-Supervisioncited by 7
  10. 2025Interpretable Hierarchical Concept Reasoning through Attention-Guided Graph Learningcited by 3
  11. 2022NESTER: An Adaptive Neurosymbolic Method for Causal Effect Estimationcited by 1
  12. 2026Grounding vs. Compositionality: On the Non-Complementarity of Reasoning in Neuro-Symbolic Systemscited by 0
1 more
  1. 2026Concise and Logically Consistent Conformal Sets for Neuro-Symbolic Concept-Based Modelscited by 0