Modular design patterns for hybrid learning and reasoning systems

Bekkum, Michael van; Boer, Maaike de; Harmelen, Frank van; Meyer-Vitali, André; Teije, Annette ten

doi:10.1007/s10489-021-02394-3

Cited by 50 publications

(13 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to further formalise the framework, we will refer to the modular design patterns and taxonomical vocabulary introduced by Bekkum et al [17], which were proposed in order to describe in a unifying way the architecture of a whole range of systems that combine statistical and symbolical methods. This means it provides the vocabulary to describe this monitoring setting in an abstract framework.…”

Section: Framework Design Patternmentioning

confidence: 99%

Monitoring AI systems: A Problem Analysis, Framework and Outlook

Onnes

2022

Preprint

View full text Add to dashboard Cite

Knowledge-based systems have been used to monitor machines and processes in the real world. In this paper we propose the use of knowledge-based systems to monitor other AI systems in operation. We motivate and provide a problem analysis of this novel setting and subsequently propose a framework that allows for structuring future research related to this setting. Several directions for further research are also discussed.

show abstract

Section: Framework Design Patternmentioning

confidence: 99%

Monitoring AI systems: A Problem Analysis, Framework and Outlook

Onnes

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Research on how to combine connectionist and symbolic approaches has flourished in the past few years [5], [12], with several applications in semantic image interpretation and visual query answering [5], [4], [13], [3], [14], [15], [16]. Among the plethora of compositional patterns that have been proposed [17], [12], the present work follows two main principles: knowledge representation (in the form of first order logic) is embedded into a neural network, which in turn allows to constrain the search space by leveraging explicit (and human-interpretable) domain knowledge as a symbolic prior. This latter property is extremely useful in ZSL, in which some external source of information is exploited to offer an abstract description of the classes in lieu of providing training examples.…”

Section: A Neural-symbolic Ai In Semantic Image Interpretationmentioning

confidence: 99%

PROTOtypical Logic Tensor Networks (PROTO-LTN) for Zero Shot Learning

Martone¹,

Manigrasso²,

Fabrizio³

et al. 2022

Preprint

View full text Add to dashboard Cite

Semantic image interpretation can vastly benefit from approaches that combine sub-symbolic distributed representation learning with the capability to reason at a higher level of abstraction. Logic Tensor Networks (LTNs) are a class of neuro-symbolic systems based on a differentiable, first-order logic grounded into a deep neural network. LTNs replace the classical concept of training set with a knowledge base of fuzzy logical axioms. By defining a set of differentiable operators to approximate the role of connectives, predicates, functions and quantifiers, a loss function is automatically specified so that LTNs can learn to satisfy the knowledge base. We focus here on the subsumption or isOfClass predicate, which is fundamental to encode most semantic image interpretation tasks. Unlike conventional LTNs, which rely on a separate predicate for each class (e.g., dog, cat), each with its own set of learnable weights, we propose a common isOfClass predicate, whose level of truth is a function of the distance between an object embedding and the corresponding class prototype. The PROTOtypical Logic Tensor Networks (PROTO-LTN) extend the current formulation by grounding abstract concepts as parametrized class prototypes in a high-dimensional embedding space, while reducing the number of parameters required to ground the knowledge base. We show how this architecture can be effectively trained in the few and zero-shot learning scenarios. Experiments on Generalized Zero Shot Learning benchmarks validate the proposed implementation as a competitive alternative to traditional embedding-based approaches. The proposed formulation opens up new opportunities in zero shot learning settings, as the LTN formalism allows to integrate background knowledge in the form of logical axioms to compensate for the lack of labelled examples.

show abstract

“…More recently, [59] analysed the intersection between NeSy and graph neural networks (GNN). [105] described neural symbolic systems in terms of the composition of blocks described by few patterns, concerning processes and exchanged data. In contrast, this survey is more focused on the underlying principles that govern such a composition.…”

Section: And G Marcus )mentioning

confidence: 99%

From Statistical Relational to Neural Symbolic Artificial Intelligence: a Survey

Marra¹,

Dumančić²,

Manhaeve³

et al. 2021

Preprint

View full text Add to dashboard Cite

Neural-symbolic and statistical relational artificial intelligence both integrate frameworks for learning with logical reasoning. This survey identifies several parallels across seven different dimensions between these two fields. These cannot only be used to characterize and position neuralsymbolic artificial intelligence approaches but also to identify a number of directions for further research.

show abstract

Modular design patterns for hybrid learning and reasoning systems

Cited by 50 publications

References 41 publications

Monitoring AI systems: A Problem Analysis, Framework and Outlook

Monitoring AI systems: A Problem Analysis, Framework and Outlook

PROTOtypical Logic Tensor Networks (PROTO-LTN) for Zero Shot Learning

From Statistical Relational to Neural Symbolic Artificial Intelligence: a Survey

Contact Info

Product

Resources

About