Train and Test Tightness of LP Relaxations in Structured Prediction
نویسندگان
چکیده
Structured prediction is used in areas such as computer vision and natural language processing to predict structured outputs such as segmentations or parse trees. In these settings, prediction is performed by MAP inference or, equivalently, by solving an integer linear program. Because of the complex scoring functions required to obtain accurate predictions, both learning and inference typically require the use of approximate solvers. We propose a theoretical explanation to the striking observation that approximations based on linear programming (LP) relaxations are often tight on real-world instances. In particular, we show that learning with LP relaxed inference encourages integrality of training instances, and that tightness generalizes from train to test data.
منابع مشابه
On the Tightness of LP Relaxations for Structured Prediction
Structured prediction applications often involve complex inference problems that require the use of approximate methods. Approximations based on linear programming (LP) relaxations have proved particularly successful in this setting, with both theoretical and empirical support. Despite the general intractability of inference, it has been observed that in many real-world applications the LP rela...
متن کاملConditions beyond treewidth for tightness of higher-order LP relaxations
Linear programming (LP) relaxations are a popular method to attempt to find a most likely configuration of a discrete graphical model. If a solution to the relaxed problem is obtained at an integral vertex then the solution is guaranteed to be exact and we say that the relaxation is tight. We consider binary pairwise models and introduce new methods which allow us to demonstrate refined conditi...
متن کاملBounding the Integrality Distance of LP Relaxations for Structured Prediction
In structured prediction, a predictor optimizes an objective function over a combinatorial search space, such as the set of all image segmentations, or the set of all part-of-speech taggings. Unfortunately, finding the optimal structured labeling—sometimes referred to as maximum a posteriori (MAP) inference—is, in general, NP-hard [12], due to the combinatorial structure of the problem. Many in...
متن کاملCharacterizing Tightness of LP Relaxations by Forbidding Signed Minors
We consider binary pairwise graphical models and provide an exact characterization (necessary and sufficient conditions observing signs of potentials) of tightness for the LP relaxation on the triplet-consistent polytope of the MAP inference problem, by forbidding an odd-K5 (complete graph on 5 variables with all edges repulsive) as a signed minor in the signed suspension graph. This captures s...
متن کاملThe More the Merrier: Parameter Learning for Graphical Models with Multiple MAPs
Conditional random field (CRFs) is a popular and effective approach to structured prediction. When the underlying structure does not have a small tree-width, maximum likelihood estimation (MLE) is in general computationally hard. Discriminative methods such as Perceptron or Max-Margin Markov Networks circumvent this problem by requiring the MAP assignment only, which is often more tractable, ei...
متن کامل