mlsanityData to 7 Oct 2026

Learning to Make Analogies by Contrasting Abstract Relational Structure

arXiv 1902.00120 ↗First posted 31 Jan 2019cs.AIcited by 53

Not on the map. Why?

Cited by 20 in LLM reasoning, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Analogical reasoning has been a principal focus of various waves of AI research. Analogy is particularly challenging for machines because it requires relational structures to be represented such that they can be flexibly applied across diverse domains of experience. Here, we study how analogical reasoning can be induced in neural networks that learn to perceive and reason about raw visual data. We find that the critical factor for inducing such a capacity is not an elaborate architecture, but rather, careful attention to the choice of data and the manner in which it is presented to the model. The most robust capacity for analogical reasoning is induced when networks learn analogies by contrasting abstract relational structures in their input domains, a training method that uses only the input data to force models to learn about important abstract features. Using this technique we demonstrate capacities for complex, visual and symbolic analogy making and generalisation in even the simplest neural network architectures.

Key facts. Cited by 53 papers in the map (data through 7 Oct 2026). First posted on arXiv 31 Jan 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 10; 2022: 9; 2023: 12; 2024: 9; 2025: 8; 2026: 5.

53 papers.

Sort:

By area:

  1. 2022Visual Prompting via Image Inpaintingcited by 248
  2. 2022Emergent Analogical Reasoning in Large Language Modelscited by 239
  3. 2021Abstraction and Analogy-Making in Artificial Intelligencecited by 103
  4. 2021Abstract Spatial-Temporal Reasoning via Probabilistic Abduction and Executioncited by 43
  5. 2022Deep Learning Methods for Abstract Visual Reasoning: A Survey on Raven's Progressive Matricescited by 37
  6. 2022A Review of Emerging Research Directions in Abstract Visual Reasoningcited by 33
  7. 2024Reasoning Abilities of Large Language Models: In-Depth Analysis on the Abstraction and Reasoning Corpuscited by 32
  8. 2021ACRE: Abstract Causal REasoning Beyond Covariationcited by 30
  9. 2024MARVEL: Multidimensional Abstraction and Reasoning through Visual Evaluation and Learningcited by 30
  10. 2025MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Modelscited by 28
  11. 2021Learning Algebraic Representation for Systematic Generalization in Abstract Reasoningcited by 27
  12. 2023In-Context Analogical Reasoning with Pre-Trained Language Modelscited by 22
38 more
  1. 2022Multimodal Analogical Reasoning over Knowledge Graphscited by 21
  2. 2023Learning to reason over visual objectscited by 19
  3. 2024The Curious Case of Nonverbal Abstract Reasoning with Multi-Modal Large Language Modelscited by 17
  4. 2021Probabilistic Analogical Mapping with Semantic Relation Networkscited by 14
  5. 2021Selective Replay Enhances Learning in Online Continual Analogical Reasoningcited by 12
  6. 2021Unsupervised Abstract Reasoning for Raven's Problem Matricescited by 12
  7. 2021Visual analogy: Deep learning versus compositional modelscited by 11
  8. 2025Learning Task Representations from In-Context Learningcited by 11
  9. 2025Analogical Reasoning Inside Large Language Models: Concept Vectors and the Limits of Abstractioncited by 11
  10. 2024Mars: Situated Inductive Reasoning in an Open-World Environmentcited by 10
  11. 2022Interactive Visual Reasoning under Uncertaintycited by 8
  12. 2022Zero-shot visual reasoning through probabilistic analogical mappingcited by 8
  13. 2024Towards Compositionality in Concept Learningcited by 7
  14. 2024ARCLE: The Abstraction and Reasoning Corpus Learning Environment for Reinforcement Learningcited by 7
  15. 2025Bongard-RWR+: Real-World Representations of Fine-Grained Concepts in Bongard Problemscited by 7
  16. 2026Causality neq Invariance: Function and Concept Vectors in LLMscited by 7
  17. 2022Geoclidean: Few-Shot Generalization in Euclidean Geometrycited by 6
  18. 2023Grounding and Distinguishing Conceptual Vocabulary Through Similarity Learning in Embodied Simulationscited by 6
  19. 2023Imagine the Unseen World: A Benchmark for Systematic Generalization in Visual World Modelscited by 6
  20. 2023One Self-Configurable Model to Solve Many Abstract Visual Reasoning Problemscited by 6
  21. 2023Towards Combinatorial Generalization for Catalysts: A Kohn-Sham Charge-Density Approachcited by 5
  22. 2024Disentangling and Integrating Relational and Sensory Information in Transformer Architecturescited by 5
  23. 2023Learning Multi-Object Positional Relationships via Emergent Communicationcited by 4
  24. 2023Im-Promptu: In-Context Composition from Image Promptscited by 4
  25. 2024A-I-RAVEN and I-RAVEN-Mesh: Two New Benchmarks for Abstract Visual Reasoningcited by 4
  26. 2021Projection: A Mechanism for Human-like Reasoning in Artificial Intelligencecited by 2
  27. 2023Computational Models of Solving Raven's Progressive Matrices: A Comprehensive Introductioncited by 2
  28. 2025Systematic Abductive Reasoning via Diverse Relation Representations in Vector-symbolic Architecturecited by 1
  29. 2026Implicit Rule Induction with Test-Time Task Embeddings in ARC-like Taskscited by 1
  30. 2021Toward Building Science Discovery Machinescited by 0
  31. 2022Zero-Episode Few-Shot Contrastive Predictive Coding: Solving intelligence tests without prior trainingcited by 0
  32. 2023Dissecting Continual Learning a Structural and Data Analysiscited by 0
  33. 2023Deep Non-Monotonic Reasoning for Visual Abstract Reasoning Taskscited by 0
  34. 2023Determinantal Point Process Attention Over Grid Cell Code Supports Out of Distribution Generalizationcited by 0
  35. 2024Two pathways to resolve relational inconsistenciescited by 0
  36. 2025GenVP: Generating Visual Puzzles with Contrastive Hierarchical VAEscited by 0
  37. 2025Advancing Generalization Across a Variety of Abstract Visual Reasoning Taskscited by 0
  38. 2025Beyond Task-Specific Reasoning: A Unified Conditional Generative Framework for Abstract Visual Reasoningcited by 0

50 most cited of 53 All 53 (JSON)