Scaling up Inference in MLNs with Spark

Islam, Mohammad Maminur; Farabi, Khan Mohammad Al; Sarkhel, Somdeb; Venugopal, Deepak

doi:10.1109/bigdata.2018.8622607

Cited by 1 publication

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where n i (⋅, ⋅) is the satisfaction of the i th ground rule. Since the potentials measure the satisfaction of the ground rules, the ŝ is set to one and the MAP inference objective in MLNs can be expressed as a weighted count of the number of satisfied ground rules: Many implementations of MLNs exist (Noessner et al 2013;Shavlik and Natarajan 2009;Niu et al 2011;Venugopal et al 2016;Islam et al 2018). In this paper we are interested in understanding how different weight learning techniques effect the quality of predictions, rather than the efficiency of inference.…”

Section: Markov Logic Networkmentioning

confidence: 99%

A taxonomy of weight learning methods for statistical relational learning

et al. 2021

View full text Add to dashboard Cite

Statistical relational learning (SRL) frameworks are effective at defining probabilistic models over complex relational data. They often use weighted first-order logical rules where the weights of the rules govern probabilistic interactions and are usually learned from data. Existing weight learning approaches typically attempt to learn a set of weights that maximizes some function of data likelihood; however, this does not always translate to optimal performance on a desired domain metric, such as accuracy or F1 score. In this paper, we introduce a taxonomy of search-based weight learning approaches for SRL frameworks that directly optimize weights on a chosen domain performance metric. To effectively apply these search-based approaches, we introduce a novel projection, referred to as scaled space (SS), that is an accurate representation of the true weight space. We show that SS removes redundancies in the weight space and captures the semantic distance between the possible weight configurations. In order to improve the efficiency of search, we also introduce an approximation of SS which simplifies the process of sampling weight configurations. We demonstrate these approaches on two state-of-the-art SRL frameworks: Markov logic networks and probabilistic soft logic. We perform empirical evaluation on five real-world datasets and evaluate them each on two different metrics. We also compare them against four other weight learning approaches. Our experimental results show that our proposed search-based approaches outperform likelihood-based approaches and yield up to a 10% improvement across a variety of performance metrics. Further, we perform an extensive evaluation to measure the robustness of our approach to different initializations and hyperparameters. The results indicate that our approach is both accurate and robust.

show abstract

Section: Markov Logic Networkmentioning

confidence: 99%