Work

Structured Prediction Models via the Matrix-Tree Theorem

Year: 2007

Type: conference-paper

Abstract: This paper provides an algorithmic framework for learning statistical models involving directed spanning trees, or equivalently non-projective dependency structures. We show how partition functions and marginals for directed spanning trees can be computed by an adaptation of Kirchhoff’s Matrix-Tree Theorem. To demonstrate an application of the method, we perform experiments which use the algorithm in training both log-linear and max-margin dependency parsers. The new training methods give improvements in accuracy over perceptron-trained models. 1

Source: RECERCAT (Consorci de Serveis Universitaris de Catalunya)

Authors: Terry Koo, Amir Globerson, Xavier Carreras, Michael Collins

Institution: Massachusetts Institute of Technology

Language: English

Topic: Natural Language Processing Techniques

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funder: National Science Foundation

Awards: MSPA-MCS: collaborative research: Statistical Learning Methods for Complex Decision Problems in Natural Language Processing, CAREER: Statistical Learning Theory for Natural Language Processing: Theory, Algorithms and Representations

OpenAlex ID: W1574662932

FWCI: 4.76

Cites: 32

Cited by: 121

Related to: 20