@article {bac:jmlr17, title = {Hinge-Loss Markov Random Fields and Probabilistic Soft Logic}, journal = {Journal of Machine Learning Research (JMLR)}, volume = {18}, year = {2017}, pages = {1-67}, abstract = {

A fundamental challenge in developing high-impact machine learning technologies is balancing the need to model rich, structured domains with the ability to scale to big data. Many important problem areas are both richly structured and large scale, from social and biological networks, to knowledge graphs and the Web, to images, video, and natural language. In this paper, we introduce two new formalisms for modeling structured data, and show that they can both capture rich structure and scale to big data. The first, hingeloss Markov random fields (HL-MRFs), is a new kind of probabilistic graphical model that generalizes different approaches to convex inference. We unite three approaches from the randomized algorithms, probabilistic graphical models, and fuzzy logic communities, showing that all three lead to the same inference objective. We then define HL-MRFs by generalizing this unified objective. The second new formalism, probabilistic soft logic (PSL), is a probabilistic programming language that makes HL-MRFs easy to define using a syntax based on first-order logic. We introduce an algorithm for inferring most-probable variable assignments (MAP inference) that is much more scalable than general-purpose convex optimization methods, because it uses message passing to take advantage of sparse dependency structures. We then show how to learn the parameters of HL-MRFs. The learned HL-MRFs are as accurate as analogous discrete models, but much more scalable. Together, these algorithms enable HL-MRFs and PSL to model rich, structured data at scales not previously possible.

}, url = {https://github.com/stephenbach/bach-jmlr17-code}, author = {Bach, Stephen H. and Broecheler, Matthias and Huang, Bert and Lise Getoor} } @article {farnadi:mlj17, title = {Soft quantification in statistical relational learning}, journal = {Machine Learning Journal}, year = {2017}, author = {Golnoosh Farnadi and Bach, Stephen H. and Moens, Marie-Francine and Lise Getoor and De Cock, Martine} } @article {bach:arxiv15, title = {Hinge-Loss Markov Random Fields and Probabilistic Soft Logic}, journal = {ArXiv:1505.04406 [cs.LG]}, year = {2015}, note = {To reference this work, please cite the JMLR paper.}, author = {Bach, Stephen H. and Broecheler, Matthias and Huang, Bert and Lise Getoor} } @mastersthesis {bach:thesis15, title = {Hinge-Loss Markov Random Fields and Probabilistic Soft Logic: A Scalable Approach to Structured Prediction}, year = {2015}, note = {Winner of the Larry S. Davis Doctoral Dissertation Award}, school = {University of Maryland, College Park}, type = {phd}, author = {Bach, Stephen H.} } @conference {bach:icml15, title = {Paired-Dual Learning for Fast Training of Latent Variable Hinge-Loss MRFs}, booktitle = {International Conference on Machine Learning (ICML)}, year = {2015}, note = {Stephen Bach and Bert Huang contributed equally.}, abstract = {

Latent variables allow probabilistic graphical models to capture nuance and structure in important domains such as network science, natural language processing, and computer vision. Naive approaches to learning such complex models can be prohibitively expensive{\textemdash}because they require repeated inferences to update beliefs about latent variables{\textemdash}so lifting this restriction for useful classes of models is an important problem. Hinge-loss Markov random fields (HL-MRFs) are graphical models that allow highly scalable inference and learning in structured domains, in part by representing structured problems with continuous variables. However, this representation leads to challenges when learning with latent variables. We introduce paired-dual learning, a framework that greatly speeds up training by using tractable entropy surrogates and avoiding repeated inferences. Paired-dual learning optimizes an objective with a pair of dual inference problems. This allows fast, joint optimization of parameters and dual variables. We evaluate on social-group detection, trust prediction in social networks, and image reconstruction, finding that paired-dual learning trains models as accurate as those trained by traditional methods in much less time, often before traditional methods make even a single parameter update.

}, author = {Bach, Stephen H. and Huang, Bert and Boyd-Graber, Jordan and Lise Getoor} } @conference {farnadi:ilp15, title = {Statistical Relational Learning with Soft Quantifiers}, booktitle = {International Conference on Inductive Logic Programming (ILP)}, year = {2015}, note = {Winner of Best Student Paper award.}, author = {Golnoosh Farnadi and Bach, Stephen H. and Blondeel, Marjon and Moens, Marie-Francine and Lise Getoor and De Cock, Martine} } @conference {bach:aistats15, title = {Unifying Local Consistency and MAX SAT Relaxations for Scalable Inference with Rounding Guarantees}, booktitle = {Artificial Intelligence and Statistics (AISTATS)}, year = {2015}, author = {Bach, Stephen H. and Huang, Bert and Lise Getoor} } @conference {farnadi:starai14, title = {Extending PSL with Fuzzy Quantifiers}, booktitle = {International Workshop on Statistical Relational Artificial Intelligence (StaRAI)}, year = {2014}, author = {Golnoosh Farnadi and Bach, Stephen H. and Moens, Marie-Francine and Lise Getoor and De Cock, Martine} } @conference {bach:dssg14, title = {Probabilistic Soft Logic for Social Good}, booktitle = {KDD Workshop on Data Science for Social Good}, year = {2014}, author = {Bach, Stephen H. and Huang, Bert and Lise Getoor} } @conference {bach:discml14, title = {Rounding Guarantees for Message-Passing MAP Inference with Logical Dependencies}, booktitle = {NIPS Workshop on Discrete and Combinatorial Problems in Machine Learning (DISCML)}, year = {2014}, author = {Bach, Stephen H. and Huang, Bert and Lise Getoor} } @conference {bach:uai13, title = {Hinge-loss Markov Random Fields: Convex Inference for Structured Prediction}, booktitle = {Uncertainty in Artificial Intelligence}, year = {2013}, abstract = {

Graphical models for structured domains are powerful tools, but the computational complexities of combinatorial prediction spaces can force restrictions on models, or require approximate inference in order to be tractable. Instead of working in a combinatorial space, we use hinge-loss Markov random fields (HL-MRFs), an expressive class of graphical models with log-concave density functions over continuous variables, which can represent confidences in discrete predictions. This paper demonstrates that HLMRFs are general tools for fast and accurate structured prediction. We introduce the first inference algorithm that is both scalable and applicable to the full class of HL-MRFs, and show how to train HL-MRFs with several learning algorithms. Our experiments show that HL-MRFs match or surpass the predictive performance of state-of-the-art methods, including discrete models, in four application domains.

}, author = {Bach, Stephen H. and Huang, Bert and London, Ben and Lise Getoor} } @conference {bach:fna13, title = {Large-margin Structured Learning for Link Ranking}, booktitle = {NIPS Workshop on Frontiers of Network Analysis: Methods, Models, and Applications}, year = {2013}, note = {Winner of Best Student Paper award}, author = {Bach, Stephen H. and Huang, Bert and Lise Getoor} } @conference {bach:inferning13, title = {Learning Latent Groups with Hinge-loss Markov Random Fields}, booktitle = {ICML Workshop on Inferning: Interactions between Inference and Learning}, year = {2013}, author = {Bach, Stephen H. and Huang, Bert and Lise Getoor} } @conference {memory:ursw12, title = {Graph Summarization in Annotated Data Using Probabilistic Soft Logic}, booktitle = {Proceedings of the International Workshop on Uncertainty Reasoning for the Semantic Web (URSW)}, year = {2012}, author = {Memory, Alex and Kimmig, Angelika and Bach, Stephen H. and Raschid, Louiqa and Lise Getoor} } @conference {bach:pmpm10, title = {Decision-Driven Models with Probabilistic Soft Logic}, booktitle = {NIPS Workshop on Predictive Models in Personalized Medicine}, year = {2010}, author = {Bach, Stephen H. and Broecheler, Matthias and Kok, Stanley and Lise Getoor} }