2009 PredictingStructureObjectsWithSVMs
- (Joachims et al., 2009) ⇒ Thorsten Joachims, Thomas Hofmann, Yisong Yue, and Chun-Nam Yu. (2009). “Predicting Structured Objects with Support Vector Machines.” In: Communications of the ACM, 52(11). doi:10.1145/1592761.1592783.
Subject Headings: Complex Object Prediction Task, Structured SVM Algorithm.
Notes
- It proposes an algorithm that incorporates structure through representations similar to those of probabilistic graphical models, such as Markov Random Fields.
- It proposes a training algorithm based on the optimization of discriminative measures of performance.
- The Optimization Task involves an exponentially large number of constraints in the problem size.
- It proposes a Cutting-Plane Algorithm to consider a restricted set of constraints.
Quotes
Abstract
Machine Learning today offers a broad repertoire of methods for classification and regression. But what if we need to predict complex objects like trees, orderings, or alignments? Such problems arise naturally in natural language processing, search engines, and bioinformatics. The following explores a generalization of Support Vector Machines (SVMs) for such complex prediction problems.
1. Introduction
…
Obviously, this question arises not only for learning to predict trees, but similarly for a variety of other structured and complex outputs. Structured output prediction is the name for such learning tasks, where one aims at learning a function h: X → Y mapping inputs x isin [math]\displaystyle{ X }[/math] to complex and structured outputs y isin Y.
…
,
| Author | volume | Date Value | title | type | journal | titleUrl | doi | note | year | |
|---|---|---|---|---|---|---|---|---|---|---|
| 2009 PredictingStructureObjectsWithSVMs | Thorsten Joachims Thomas Hofmann Yisong Yue Chun-Nam Yu | Predicting Structured Objects with Support Vector Machines | Communications of the ACM | 10.1145/1592761.1592783 | 2009 |