SAFRAN: An interpretable, rule-based link prediction method outperforming embedding models

Ott, Simon; Meilicke, Christian; Samwald, Matthias

doi:10.48550/arxiv.2109.08002

Cited by 5 publications

(9 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, one could refer back to the rules which governed the algorithm to get a humanreadable understanding of how and why certain predictions were made. SAFRAN [112], for example, which was inspired by AnyBURL [98], provides post-hoc explanations based on rules for every prediction. When predicting the genres of music by Bryan Adams, SAFRAN uses the following rule to incorrectly predict Heavy Metal based on the fact that Adams also does Hard Rock: genre(HeavyM etal, Y ) ← genre(HardRock, Y ) However, SAFRAN assigns confidences to soft rules, so we could see that this rule was given too high a confidence.…”

Section: B Rule-based Methods For Knowledge Graph Completionmentioning

confidence: 99%

“…When predicting the genres of music by Bryan Adams, SAFRAN uses the following rule to incorrectly predict Heavy Metal based on the fact that Adams also does Hard Rock: genre(HeavyM etal, Y ) ← genre(HardRock, Y ) However, SAFRAN assigns confidences to soft rules, so we could see that this rule was given too high a confidence. This interpretability allows model tuning, adjustment for incorrect predictions, explanations for end-users, and more [112].…”

Section: B Rule-based Methods For Knowledge Graph Completionmentioning

confidence: 99%

See 1 more Smart Citation

Neurosymbolic AI for Reasoning on Graph Structures: A Survey

DeLong¹,

Fernández²,

Whyte³

et al. 2023

Preprint

View full text Add to dashboard Cite

Neurosymbolic AI is an increasingly active area of research which aims to combine symbolic reasoning methods with deep learning to generate models with both high predictive performance and some degree of human-level comprehensibility. As knowledge graphs are becoming a popular way to represent heterogeneous and multi-relational data, methods for reasoning on graph structures have attempted to follow this neurosymbolic paradigm. Traditionally, such approaches have utilized either rule-based inference or generated representative numerical embeddings from which patterns could be extracted. However, several recent studies have attempted to bridge this dichotomy in ways that facilitate interpretability, maintain performance, and integrate expert knowledge. Within this article, we survey a breadth of methods that perform neurosymbolic reasoning tasks on graph structures. To better compare the various methods, we propose a novel taxonomy by which we can classify them. Specifically, we propose three major categories: (1) logicallyinformed embedding approaches, (2) embedding approaches with logical constraints, and (3) rule-learning approaches. Alongside the taxonomy, we provide a tabular overview of the approaches and links to their source code, if available, for more direct comparison. Finally, we discuss the applications on which these methods were primarily used and propose several prospective directions toward which this new field of research could evolve.

show abstract

Section: B Rule-based Methods For Knowledge Graph Completionmentioning

confidence: 99%

Section: B Rule-based Methods For Knowledge Graph Completionmentioning

confidence: 99%

Neurosymbolic AI for Reasoning on Graph Structures: A Survey

DeLong¹,

Fernández²,

Whyte³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…However, its aggregation process is affected by the redundancy of multiple rules. Thus, the Scalable and Fast Non-redundant Rule Application (SAFRAN) [30] was used to detect cluster redundant rules before aggregation, thereby improving the effect of the AnyBURL model. Based on feature learning methods, in combat systems, diverse systems and information flows were treated as nodes and edges, using the Representation Learning based Heterogeneous Combat Network (RLHCN) [31] method to predict various link types in the combat network.…”

Section: Link Prediction Methodsmentioning

confidence: 99%

Sparse Subgraph Prediction Based on Adaptive Attention

Gao

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

Link prediction is a crucial problem in the analysis of graph-structured data, and graph neural networks (GNNs) have proven to be effective in addressing this problem. However, the computational and temporal costs associated with large-scale graphs remain a concern. This study introduces a novel method for link prediction called Sparse Subgraph Prediction Based on Adaptive Attention (SSP-AA). The method generates sparse subgraphs and utilizes Graph SAmple and aggreGatE (GraphSAGE) for prediction, aiming to reduce computation and time costs while providing a foundation for future exploration of large-scale graphs. Certain key issues in GraphSAGE are addressed by integrating an adaptive attention mechanism and a jumping knowledge module into the model. To address the issue of adaptive weight distribution in GraphSAGE, an aggregation function is employed, which is based on the attention mechanism. This modification enables the model to distribute weights adaptively among neighboring nodes, significantly improving its ability to capture node relationships. Furthermore, to tackle the common issue of over-smoothing in GNNs, a jumping knowledge module is integrated, enabling information sharing across different layers and providing the model with the flexibility to select the appropriate representation depth based on the specific situation. By enhancing the quality of node representations, SSP-AA further boosts the performance of GraphSAGE in various prediction tasks involving graph-structured data.

show abstract

“…This corresponds to the prediction block in Figure 1. There are three different approaches to determine the score of each candidate: Maximum score and Noisy-OR originally proposed with AnyBURL in [25], and Non-redundant Noisy-OR proposed as a framework called SAFRAN [26].…”

Section: Fixed Path Generatormentioning

confidence: 99%

“…The first aggregation evaluates a set of policies, and the second aggregation evaluates the set of paths or trajectories generated by each policy. The functions ϕ ω and ψ ω can take A A PREPRINT -SEPTEMBER 26,2023 different forms depending on the model and represent the importance of the path and the policy in the prediction. The quality of the policy µ i is represented through ψ ω , which is a trainable parameter that weights the importance of different policies.…”

Section: A Mathematical Framework Formulationmentioning

confidence: 99%

An Open Access Strategy for the Drug Repurposing Community

Dawson

2023

Preprint

View full text Add to dashboard Cite

To ensure the widest possible dissemination of research results to the academic community, pharmaceutical industry, patients and to the broader public, the EU-funded drug repurposing project REPO4EU is committed to an Open Science approach. Because Open Science can be interpreted widely, this document lays out the strategy of the project with regard to Open Access publishing, alternative metrics, intellectual property and FAIR data, in line with the goals of the European Commission. The Open Science Strategy forms the theoretical framework for the REPO4EU Open Science publishing portal that will develop into an open hub of research results and communication for the entire drug repurposing community.

show abstract

SAFRAN: An interpretable, rule-based link prediction method outperforming embedding models

Cited by 5 publications

References 12 publications

Neurosymbolic AI for Reasoning on Graph Structures: A Survey

Neurosymbolic AI for Reasoning on Graph Structures: A Survey

Sparse Subgraph Prediction Based on Adaptive Attention

An Open Access Strategy for the Drug Repurposing Community

Contact Info

Product

Resources

About