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

  title={Modular Design Patterns for Hybrid Learning and Reasoning Systems: a taxonomy, patterns and use cases},
  author={Michael A. van Bekkum and Maaike de Boer and F. V. Harmelen and Andr{\'e} Meyer-Vitali and Annette ten Teije},
  journal={Appl. Intell.},
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… 
Natlog: a Lightweight Logic Programming Language with a Neuro-symbolic Touch
  • Paul Tarau
  • Computer Science
    ICLP Technical Communications
  • 2021
Natlog, a lightweight Logic Programming language, sharing Prolog’s unificationdriven execution model, but with a simplified syntax and semantics is introduced, tightly embedded in the Python-based deep-learning ecosystem with focus on content-driven indexing of ground term datasets.
Modular design patterns for neural-symbolic integration: refinement and combination
We formalise some aspects of the neural-symbol design patterns of van Bekkum et al., such that we can formally define notions of refinement of patterns, as well as modular combination of larger
Monitoring AI systems: A Problem Analysis, Framework and Outlook
This paper aims to study how to monitor AI systems in such a way that the expectations for the performance are formulated into an interpretable, knowledge-based system for monitoring.
A Framework for Context-Dependent Augmented Reality Applications Using Machine Learning and Ontological Reasoning
A framework that combines machine learning services for object recognition with ontological reasoning for determining the current situation of a user based on sensor data and deriving according actions for information display is described.
Meaningful human control: actionable properties for AI system development
How can humans remain in control of artificial intelligence (AI)-based systems designed to perform tasks autonomously? Such systems are increasingly ubiquitous, creating benefits - but also
Alert-Driven Attack Graph Generation Using S-PDFA
This work proposes SAGE, an explainable sequence learning pipeline that automatically constructs AGs from intrusion alerts without a priori expert knowledge that provides directly relevant insights into strategic differences and fingerprintable paths.
Meaningful human control over AI systems: beyond talking the talk
This paper argues these four properties are necessary for AI systems under meaningful human control, and provides possible directions to incorporate them into practice, and illustrates these properties with two use cases, automated vehicle and AI-based hiring.
Bridging Perception, Memory, and Inference through Semantic Relations
This work probes six popular language models for semantic relations and outlines a future line of research to study how the constituent subsystems can be jointly realised and integrated.
Natural Computational Architectures for Cognitive Info-Communication
A set of perspectives and approaches which have shaped theDevelopment of biologically inspired computational models in the recent past that can lead to the development of biologically more realistic cognitive architectures are presented.
Modular Design Patterns for Hybrid Actors
An extension of the taxonomy to describe distributed hybrid AI systems with actors and interactions and examples using a few design patterns relevant in multi-agent systems and human-agent interaction in general and, specifically, in the manufacturing domain are shown.


A Boxology of Design Patterns for Hybrid Learning and Reasoning Systems
A set of compositional design patterns are proposed to describe a large variety of systems that combine statistical techniques from machine learning with symbolic techniques from knowledge representation to systematize the literature, clarify which combinations of techniques serve which purposes, and encourage re-use of software components.
What the is a Symbol
The notion of a `symbol' plays an important role in the disciplines of Philosophy, Psychology, Computer Science, and Cognitive Science. However, there is comparatively little agreement on how this
Mastering the game of Go without human knowledge
An algorithm based solely on reinforcement learning is introduced, without human data, guidance or domain knowledge beyond game rules, that achieves superhuman performance, winning 100–0 against the previously published, champion-defeating AlphaGo.
Knowledge Graph Embedding: A Survey of Approaches and Applications
This article provides a systematic review of existing techniques of Knowledge graph embedding, including not only the state-of-the-arts but also those with latest trends, based on the type of information used in the embedding task.
Human-level intelligence or animal-like abilities?
What just happened in artificial intelligence and how it is being misunderstood.
Knowledge graph refinement: A survey of approaches and evaluation methods
A survey of such knowledge graph refinement approaches, with a dual look at both the methods being proposed as well as the evaluation methodologies used.
Ontology learning from text: A look back and into the future
This survey looks at how far the authors have come since the turn of the millennium and discusses the remaining challenges that will define the research directions in this area in the near future.
A Review of Relational Machine Learning for Knowledge Graphs
This paper provides a review of how statistical models can be “trained” on large knowledge graphs, and then used to predict new facts about the world (which is equivalent to predicting new edges in the graph) and how such statistical models of graphs can be combined with text-based information extraction methods for automatically constructing knowledge graphs from the Web.
Gradient-based learning applied to document recognition
This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task, and Convolutional neural networks are shown to outperform all other techniques.