Sciweavers

Share
JMLR
2012

Deterministic Annealing for Semi-Supervised Structured Output Learning

7 years 6 months ago
Deterministic Annealing for Semi-Supervised Structured Output Learning
In this paper we propose a new approach for semi-supervised structured output learning. Our approach uses relaxed labeling on unlabeled data to deal with the combinatorial nature of the label space and further uses domain constraints to guide the learning. Since the overall objective is non-convex, we alternate between the optimization of the model parameters and the label distribution of unlabeled data. The alternating optimization coupled with deterministic annealing helps us achieve better local optima and as a result our approach leads to better constraint satisfaction during inference. Experimental results on sequence labeling benchmarks show superior performance of our approach compared to Constraint Driven Learning (CoDL) and Posterior Regularization (PR).
Paramveer S. Dhillon, S. Sathiya Keerthi, Kedar Be
Added 27 Sep 2012
Updated 27 Sep 2012
Type Journal
Year 2012
Where JMLR
Authors Paramveer S. Dhillon, S. Sathiya Keerthi, Kedar Bellare, Olivier Chapelle, Sundararajan Sellamanickam
Comments (0)
books