Neural Networks Enhancement with Logical Knowledge

Daniele, Alessandro; Serafini, Luciano

doi:10.48550/arxiv.2009.06087

Cited by 2 publications

(7 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To refine the predictions of the GNN, one or more knowledge enhancement layers are stacked on top of it with the goal to increase the satisfaction of the prior knowledge. The layers consist of clause enhancers for each clause in the prior knowledge that implement a so-called differentiable boost function [15]. It returns adjustments for the GNN's predictions with respect to a clause.…”

Section: Symbolic Componentmentioning

confidence: 99%

“…The changes are computed with the objective to increase the value of the t-conorm ⊥(𝑡), which measures the satisfaction of the clause. The following function [15] is used as a differentiable approximate for the Gödel t-conorm boost function:…”

Section: Symbolic Componentmentioning

confidence: 99%

“…The clause weights obtained during training provide insights on the extent of adjustment applied by a clause. As a reference, the clause compliance (see Appendix B) [15] measures the satisfaction of a clause in the training set before optimization. Fig.…”

Section: Clause Weight Analysismentioning

confidence: 99%

“…The clause compliance [15] indicates the level of satisfaction of a clause in the data in this experimental setting. Given a clause 𝑐, a class 𝑘, the set of training nodes V train , the set of nodes of the class 𝑘: V 𝑘 = {𝑣|𝑣 ∈ V train ∧ Cls(𝑣) == 𝑘}, and the neighborhood 𝒩 (𝑣) of 𝑣, the clause compliance on graph G is defined as follows:…”

Section: B Clause Weight Evaluationmentioning

confidence: 99%

“…In the multifaceted neuro-symbolic field, KeGNN can be placed in the category of knowledgeguided learning [15], where the focus lies on learning in the presence of additional supervision induced by prior knowledge. Among these approaches, KeGNN belongs to model-based approaches where prior knowledge is an integral part of the model in the form of knowledge enhancement layers.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Knowledge Enhanced Graph Neural Networks

Werner¹,

Layaïda²,

Genevès³

et al. 2023

Preprint

View full text Add to dashboard Cite

Graph data is omnipresent and has a large variety of applications such as natural science, social networks or semantic web. Though rich in information, graphs are often noisy and incomplete. Therefore, graph completion tasks such as node classification or link prediction have gained attention. On the one hand, neural methods such as graph neural networks have proven to be robust tools for learning rich representations of noisy graphs. On the other hand, symbolic methods enable exact reasoning on graphs. We propose KeGNN, a neuro-symbolic framework for learning on graph data that combines both paradigms and allows for the integration of prior knowledge into a graph neural network model. In essence, KeGNN consists of a graph neural network as a base on which knowledge enhancement layers are stacked with the objective of refining predictions with respect to prior knowledge. We instantiate KeGNN in conjunction with two standard graph neural networks: Graph Convolutional Networks and Graph Attention Networks, and evaluate KeGNN on multiple benchmark datasets for node classification.

show abstract

Section: Symbolic Componentmentioning

confidence: 99%

Section: Symbolic Componentmentioning

confidence: 99%

Section: Clause Weight Analysismentioning

confidence: 99%

Section: B Clause Weight Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Knowledge Enhanced Graph Neural Networks

Werner¹,

Layaïda²,

Genevès³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

Logic Tensor Networks for Top-N Recommendation

Carraro

Daniele

Aiolli

et al. 2023

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Despite being studied for more than twenty years, state-of-the-art recommendation systems still suffer from important drawbacks which limit their usage in real-world scenarios. Among the well-known issues of recommender systems, there are data sparsity and the cold-start problem. These limitations can be addressed by providing some background knowledge to the model to compensate for the scarcity of data. Following this intuition, we propose to use Logic Tensor Networks (LTNs) to tackle the top-n item recommendation problem. In particular, we show how LTNs can be used to easily and effectively inject commonsense recommendation knowledge inside a recommender system. We evaluate our method on MindReader, a knowledge graph-based movie recommendation dataset containing plentiful side information. In particular, we perform an experiment to show how the benefits of the knowledge increase with the sparsity of the dataset. Eventually, a comparison with a standard Matrix Factorization approach reveals that our model is able to reach and, in many cases, outperform state-of-the-art performance.

show abstract

Neural Networks Enhancement with Logical Knowledge

Cited by 2 publications

References 0 publications

Knowledge Enhanced Graph Neural Networks

Knowledge Enhanced Graph Neural Networks

Logic Tensor Networks for Top-N Recommendation

Contact Info

Product

Resources

About