Sarawagi, Sunita ; Gupta, Rahul (2008) Accurate max-margin training for structured output spaces In: 25th international conference on Machine learning.
![]() |
PDF
430kB |
Official URL: http://doi.org/10.1145/1390156.1390268
Related URL: http://dx.doi.org/10.1145/1390156.1390268
Abstract
Tsochantaridis et al. (2005) proposed two formulations for maximum margin training of structured spaces: margin scaling and slack scaling. While margin scaling has been extensively used since it requires the same kind of MAP inference as normal structured prediction, slack scaling is believed to be more accurate and better-behaved. We present an efficient variational approximation to the slack scaling method that solves its inference bottleneck while retaining its accuracy advantage over margin scaling. We further argue that existing scaling approaches do not separate the true labeling comprehensively while generating violating constraints. We propose a new max-margin trainer PosLearn that generates violators to ensure separation at each position of a decomposable loss function. Empirical results on real datasets illustrate that PosLearn can reduce test error by up to 25% over margin scaling and 10% over slack scaling. Further, PosLearn violators can be generated more efficiently than slack violators; for many structured tasks the time required is just twice that of MAP inference.
Item Type: | Conference or Workshop Item (Paper) |
---|---|
Source: | Copyright of this article belongs to ACM, Inc |
ID Code: | 128388 |
Deposited On: | 20 Oct 2022 04:16 |
Last Modified: | 14 Nov 2022 11:09 |
Repository Staff Only: item control page