mlsanityData to 7 Oct 2026

Modular Design Patterns for Hybrid Learning and Reasoning Systems: a taxonomy, patterns and use cases

arXiv 2102.11965 ↗First posted 23 Feb 2021cs.AI cs.LGcited by 39

Not on the map. Why?

Abstract (arXiv)

The unification of statistical (data-driven) and symbolic (knowledge-driven) methods is widely recognised as one of the key challenges of modern AI. Recent years have seen large number of publications on such hybrid neuro-symbolic AI systems. That rapidly growing literature is highly diverse and mostly empirical, and is lacking a unifying view of the large variety of these hybrid systems. In this paper we analyse a large body of recent literature and we propose a set of modular design patterns for such hybrid, neuro-symbolic systems. We are able to describe the architecture of a very large number of hybrid systems by composing only a small set of elementary patterns as building blocks. The main contributions of this paper are: 1) a taxonomically organised vocabulary to describe both processes and data structures used in hybrid systems; 2) a set of 15+ design patterns for hybrid AI systems, organised in a set of elementary patterns and a set of compositional patterns; 3) an application of these design patterns in two realistic use-cases for hybrid AI systems. Our patterns reveal similarities between systems that were not recognised until now. Finally, our design patterns extend and refine Kautz' earlier attempt at categorising neuro-symbolic architectures.

Key facts. Cited by 39 papers in the map (data through 7 Oct 2026). Builds on 32 of its 75 references that are in the map. First posted on arXiv 23 Feb 2021.

Builds on

32 of 75 references.

References · 32

  1. 2018Graph Neural Networks: A Review of Methods and Applicationscited by 1,571
  2. 2018Relational inductive biases, deep learning, and graph networkscited by 1,341
  3. 2019Neural Probabilistic Logic Programming in DeepProbLogcited by 389
  4. 2017A Semantic Loss Function for Deep Learning with Symbolic Knowledgecited by 267
  5. 2015A Review of Relational Machine Learning for Knowledge Graphscited by 264
  6. 2020Neurosymbolic AI: The 3rd Wavecited by 253
  7. 2018Representation Tradeoffs for Hyperbolic Embeddingscited by 204
  8. 2017End-to-End Differentiable Provingcited by 159
  9. 2019Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoningcited by 149
  10. 2017Logic Tensor Networks for Semantic Image Interpretationcited by 92
  11. 2018Theoretical Impediments to Machine Learning With Seven Sparks from the Causal Revolutioncited by 66
  12. 2015Hinge-Loss Markov Random Fields and Probabilistic Soft Logiccited by 63
20 more
  1. 2019Iteratively Learning Embeddings and Rules for Knowledge Graph Reasoningcited by 63
  2. 2020From Statistical Relational to Neuro-Symbolic Artificial Intelligencecited by 63
  3. 2020Graph Kernels: State-of-the-Art and Future Challengescited by 63
  4. 2017Deep Network Guided Proof Searchcited by 47
  5. 2016Towards Deep Symbolic Reinforcement Learningcited by 42
  6. 2020Scalable Neural Methods for Reasoning With a Symbolic Knowledge Basecited by 32
  7. 2020Efficient Generation of Structured Objects with Constrained Adversarial Networkscited by 27
  8. 2019Unsupervised Grounding of Plannable First-Order Logic Representation from Imagescited by 24
  9. 2016Lifted Rule Injection for Relation Embeddingscited by 22
  10. 2019A Boxology of Design Patterns for Hybrid Learning and Reasoning Systemscited by 20
  11. 2020Relational Neural Machinescited by 19
  12. 2020Teaching the Old Dog New Tricks: Supervised Learning with Constraintscited by 16
  13. 2018Accuracy-based Curriculum Learning in Deep Reinforcement Learningcited by 15
  14. 2017Explaining Trained Neural Networks with Semantic Web Technologies: First Stepscited by 13
  15. 2018Improving Visual Relationship Detection using Semantic Modeling of Scene Descriptionscited by 13
  16. 2017Human-Level Intelligence or Animal-Like Abilities?cited by 8
  17. 2020Injecting Domain Knowledge in Neural Networks: a Controlled Experiment on a Constrained Problemcited by 7
  18. 2018Reasoning over RDF Knowledge Bases using Deep Learningcited by 4
  19. 2017Deep Learning for Ontology Reasoningcited by 3
  20. 2020IterefinE: Iterative KG Refinement Embeddings using Symbolic Knowledgecited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 5; 2022: 5; 2023: 9; 2024: 8; 2025: 7; 2026: 5.

39 papers.

Sort:

By area:

  1. 2021From Statistical Relational to Neurosymbolic Artificial Intelligence: a Surveycited by 72
  2. 2021Meaningful human control: actionable properties for AI system developmentcited by 18
  3. 2023Human-AI collaboration is not very collaborative yet: A taxonomy of interaction patterns in AI-assisted decision making from a systematic reviewcited by 18
  4. 2024Ontology Embedding: A Survey of Methods, Applications and Resourcescited by 12
  5. 2023Knowledge Engineering using Large Language Modelscited by 9
  6. 2024Unpacking Human-AI interactions: From interaction primitives to a design spacecited by 7
  7. 2025Defining neurosymbolic AIcited by 7
  8. 2021Natlog: a Lightweight Logic Programming Language with a Neuro-symbolic Touchcited by 5
  9. 2022Benchmark for Models Predicting Human Behavior in Gap Acceptance Scenarioscited by 5
  10. 2023Describing and Organizing Semantic Web and Machine Learning Systems in the SWeMLS-KGcited by 5
  11. 2023Seeing the Intangible: Survey of Image Classification into High-Level and Abstract Categoriescited by 4
  12. 2022PROTOtypical Logic Tensor Networks (PROTO-LTN) for Zero Shot Learningcited by 3
27 more
  1. 2024Combining Machine Learning and Ontology: A Systematic Literature Reviewcited by 3
  2. 2025The DeepLog Neurosymbolic Machinecited by 3
  3. 2021Modular Design Patterns for Hybrid Actorscited by 2
  4. 2023Ontology Pre-training for Poison Predictioncited by 2
  5. 2023Trust, Accountability, and Autonomy in Knowledge Graph-based AI for Self-determinationcited by 2
  6. 2024Aligning Generalisation Between Humans and Machinescited by 2
  7. 2026Knowledge Affordances for Hybrid Human-AI Information Seekingcited by 2
  8. 2022Modular design patterns for neural-symbolic integration: refinement and combinationcited by 1
  9. 2023Ontology in Hybrid Intelligence: a concise literature reviewcited by 1
  10. 2023Fuzzy Logic Visual Network (FLVN): A neuro-symbolic approach for visual features matchingcited by 1
  11. 2024Stitching Gaps: Fusing Situated Perceptual Knowledge with Vision Transformers for High-Level Image Classificationcited by 1
  12. 2024IID Relaxation by Logical Expressivity: A Research Agenda for Fitting Logics to Neurosymbolic Requirementscited by 1
  13. 2025Mutual Understanding between People and Systems via Neurosymbolic AI and Knowledge Graphscited by 1
  14. 2025DeepGraphLog for Layered Neurosymbolic AIcited by 1
  15. 2025Discovering Association Rules in High-Dimensional Small Tabular Datacited by 1
  16. 2026Neuro-symbolic Action Masking for Deep Reinforcement Learningcited by 1
  17. 2021Natural Computational Architectures for Cognitive Info-Communicationcited by 0
  18. 2022Monitoring AI systems: A Problem Analysis, Framework and Outlookcited by 0
  19. 2022Neuro-Symbolic Spatio-Temporal Reasoningcited by 0
  20. 2023Computational Natural Philosophy: A Thread from Presocratics through Turing to ChatGPTcited by 0
  21. 2024Standardizing Knowledge Engineering Practices with a Reference Architecturecited by 0
  22. 2024Enhancing Medical Learning and Reasoning Systems: A Boxology-Based Comparative Analysis of Design Patternscited by 0
  23. 2025"Stop replacing salt with sugar!'': Towards Intuitive Human-Agent Teachingcited by 0
  24. 2025Bridging the AI Trustworthiness Gap between Functions and Normscited by 0
  25. 2026Autoregressive Models for Knowledge Graph Generationcited by 0
  26. 2026From Hybrid Mechanistic--Data-Driven Modeling Toward Neuro-Symbolic AI: What, Why, and Howcited by 0
  27. 2026The RAIL Principles for Neurosymbolic AI: Reasoning, Assurances, Interfacing and Learningcited by 0