Discriminative Training of Markov Logic Networks

By Parag Singla, Pedro M. Domingos. AAAI Conference on Artificial Intelligence, 2005

309 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/ba8d2cfa13ac2e6981f06474a2c806eb4923ed56

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Hinge-Loss Markov Random Fields and Probabilistic Soft Logic (2015, methodology): “Computing the expectation is intractable, so we use a common approximation (e.g., Collins, 2002; Singla and Domingos, 2005; Poon and Domingos, 2011): the values of the potentials at the most probable setting of y with the current parameters, i.”
Learning the structure of Markov logic networks (2005, background): “Singla and Domingos (2005) proposed a discriminative approach to learning MLN weights, but did not learn MLN structure, like we do here.”
An Introduction to Conditional Random Fields (2010, background): “Markov logic networks [98, 110] use logical formulae to specify”
GMNN: Graph Markov Neural Networks (2019, unknown): “…such weaknesses, many later SRL methods employ Markov networks (e.g., conditional random fields (Lafferty et al., 2001)), and representative methods include relational Markov networks (RMN) (Taskar et al., 2002) and Markov logic networks (MLN) (Richardson & Domingos, 2006; Singla & Domingos, 2005).”
Automatically refining the wikipedia infobox ontology (2008, methodology): “Discriminative learning is used to determine the weights of formulas [30], and MC-SAT is used for inference [27].”
Learning from Imbalanced Data (2009, result): “As a result, many of the current research work in the community use PR curves for performance evaluations and comparisons [124], [125], [126], [127].”
Joint Inference in Information Extraction (2007, methodology): “In particular, we performed inference using the MC-SAT algorithm (Poon & Domingos 2006), and discriminative weight learning using the voted perceptron algorithm ( Singla & Domingos 2005 ).”
Entity Resolution with Markov Logic (2006, methodology): “While it was originally developed for the special case of hidden Markov models, and used the Viterbi algorithm to find the MAP state, Singla and Domingos [41] generalized it to MLNs by replacing Viterbi with MaxWalkSAT.”
The Alchemy System for Statistical Relational AI: User Manual (2007, methodology): “…familiar with Markov logic, we recommend that you read the papers Markov Logic Networks [7], Discriminative Training of Markov Logic Networks [9], Learning the Structure of Markov Logic Networks [3], Memory-Eﬃcient Inference in Relational Domains [10] and Sound and Eﬃcient Inference with…”
Legal requirements on explainability in machine learning (2020, methodology): “(2018) use a Markov logic network (MLN) (Singla and Domingos 2005) to predict the outcome of divorce judgments.”
The relationship between Precision-Recall and ROC curves (2006, methodology): “Looking at PR curves can expose di(cid:11)erences between algorithms that are not apparent in ROC space.”
Formalizing Visualization Design Knowledge as Constraints: Actionable and Extensible Models in Draco (2018, methodology): “The problem of finding the optimal completion of a partial specification is the same as performing maximum a posteriori (MAP) inference in the probability model [54].”
Lifted First-Order Belief Propagation (2008, methodology): “We used the MLN (formulas and weights) used by Singla and Domingos (2005) in their experiments.”
A user-guided approach to program analysis (2015, background): “It is an adaptation of [38] to our application.”
An Introduction to Conditional Random Fields for Relational Learning (2007, background): “Finally, Markov logic networks [Richardson and Domingos, 2005, Singla and Domingos, 2005] are a type of probabilistic logic in which there are parameters for each ﬁrst-order rule in a knowledge base.”
StatSnowball: a statistical approach to extracting entity relationships (2009, methodology): “Thus, we partition the ground atoms into two sets—the set of evidence atoms X and the set of query atoms Q, and define a discriminative MLN [23].”
Sound and Efficient Inference with Probabilistic and Deterministic Dependencies (2006, methodology): “We used a Markov logic network similar to that of Singla and Domingos (2005), containing formulas like: if two fields have high TF-IDF similarity, they are likely to be the same; if two fields are the same, their records are likely to be the same; if two records are the same, their fields are the…”