[DTC18a] SyMIL: MinMax Latent SVM for Weakly Labeled Data

Revue Internationale avec comité de lecture : Journal IEEE Transactions on Neural Networks and Learning Systems, 2018

Mots clés: Weakly Supervised Learning, Multiple Instance Learning, Latent SVM, Image Categorization and Pattern Recognition

Résumé: esigning powerful models able to handle weakly labeled data is a crucial problem in machine learning. In this paper, we propose a new Multiple Instance Learning (MIL) framework. Examples are represented as bags of instances, but we depart from standard MIL assumptions by introducing a symmetric strategy (SyMIL) that seeks discriminative instances in positive and negative bags. The idea is to use the instance the most distant from the hyper-plan to classify the bag. We provide a theoretical analysis featuring the generalization properties of our model. We derive a large margin formulation of our problem, which is cast as a difference of convex functions, and optimized using CCCP. We provide a primal version optimizing with stochastic sub-gradient descent and a dual version optimizing with one-slack cutting- plane. Successful experimental results are reported on standard MIL and weakly-supervised object detection datasets: SyMIL significantly outperforms competitive methods (mi/MI/Latent- SVM), and gives very competitive performance compared to state-of-the-art works. We also analyze the selected instances of symmetric and asymmetric approaches on weakly-supervised object detection and text classification tasks. Finally we show complementarity of SyMIL with recent works on learning with label proportions on standard MIL datasets.


@article {
title="{SyMIL: MinMax Latent SVM for Weakly Labeled Data}",
author="T. Durand and N. Thome and M. Cord",
journal="IEEE Transactions on Neural Networks and Learning Systems",