Induction of Interpretable Possibilistic Logic Theories from Relational\n Data

2017/05/19 by Ondřej Kuželka, Kuzelka, Ondrej, Jesse Davis +3 · 1 citation
Computer Science · #Artificial Intelligence (cs.AI) #Bayesian Modeling and Causal Inference #Data Mining Algorithms and Applications #FOS: Computer and information sciences #Semantic Web and Ontologies

paper · pdf · doi:10.48550/arxiv.1705.07095

openalex publication_date 2017/05/19 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

The field of Statistical Relational Learning (SRL) is concerned with learning\nprobabilistic models from relational data. Learned SRL models are typically\nrepresented using some kind of weighted logical formulas, which make them\nconsiderably more interpretable than those obtained by e.g. neural networks. In\npractice, however, these models are often still difficult to interpret\ncorrectly, as they can contain many formulas that interact in non-trivial ways\nand weights do not always have an intuitive meaning. To address this, we\npropose a new SRL method which uses possibilistic logic to encode relational\nmodels. Learned models are then essentially stratified classical theories,\nwhich explicitly encode what can be derived with a given level of certainty.\nCompared to Markov Logic Networks (MLNs), our method is faster and produces\nconsiderably more interpretable models.\n

Cited by

• Inductive Learning for Possibilistic Logic Programs Under Stable Models

Related

• Statistical relational learning and neuro-symbolic AI: what does first-order logic offer?
• Why deterministic logic is hard to learn but Statistical Relational Learning works
• Neural Markov Logic Networks
• Encoding Markov Logic Networks in Possibilistic Logic
• Learning Possibilistic Logic Theories from Default Rules