The Lovász Hinge: A Convex Surrogate for Submodular Losses
نویسندگان
چکیده
Learning with non-modular losses is an important problem when sets of predictions are made simultaneously. The main tools for constructing convex surrogate loss functions for set prediction are margin rescaling and slack rescaling. In this work, we show that these strategies lead to tight convex surrogates iff the underlying loss function is increasing in the number of incorrect predictions. However, gradient or cutting-plane computation for these functions is NP-hard for non-supermodular loss functions. We propose instead a novel surrogate loss function for submodular losses, the Lovász hinge, which leads to O(p log p) complexity with O(p) oracle accesses to the loss function to compute a gradient or cutting-plane. We prove that the Lovász hinge is convex and yields an extension. As a result, we have developed the first tractable convex surrogates in the literature for submodular losses. We demonstrate the utility of this novel convex surrogate through several set prediction tasks, including on the PASCAL VOC and Microsoft COCO datasets.
منابع مشابه
Learning Submodular Losses with the Lovasz Hinge
Learning with non-modular losses is an important problem when sets of predictions are made simultaneously. The main tools for constructing convex surrogate loss functions for set prediction are margin rescaling and slack rescaling. In this work, we show that these strategies lead to tight convex surrogates iff the underlying loss function is increasing in the number of incorrect predictions. Ho...
متن کاملA Convex Surrogate Operator for General Non-Modular Loss Functions
Empirical risk minimization frequently employs convex surrogates to underlying discrete loss functions in order to achieve computational tractability during optimization. However, classical convex surrogates can only tightly bound modular loss functions, submodular functions or supermodular functions separately while maintaining polynomial time computation. In this work, a novel generic convex ...
متن کاملOptimization of the Jaccard index for image segmentation with the Lovász hinge
The Jaccard loss, commonly referred to as the intersection-over-union loss, is commonly employed in the evaluation of segmentation quality due to its better perceptual quality and scale invariance, which lends appropriate relevance to small objects compared with per-pixel losses. We present a method for direct optimization of the per-image intersection-over-union loss in neural networks, in the...
متن کاملMinimizing The Misclassification Error Rate Using a Surrogate Convex Loss
We carefully study how well minimizing convex surrogate loss functions corresponds to minimizing the misclassification error rate for the problem of binary classification with linear predictors. We consider the agnostic setting, and investigate guarantees on the misclassification error of the loss-minimizer in terms of the margin error rate of the best predictor. We show that, aiming for such a...
متن کاملShaping Level Sets with Submodular Functions
We consider a class of sparsity-inducing regularization terms based on submodular functions. While previous work has focused on non-decreasing functions, we explore symmetric submodular functions and their Lovász extensions. We show that the Lovász extension may be seen as the convex envelope of a function that depends on level sets (i.e., the set of indices whose corresponding components of th...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- CoRR
دوره abs/1512.07797 شماره
صفحات -
تاریخ انتشار 2015