https://hal.inria.fr/hal-00773605Kumar, M. PawanM. PawanKumarGALEN - Organ Modeling through Extraction, Representation and Understanding of Medical Image Content - Inria Saclay - Ile de France - Inria - Institut National de Recherche en Informatique et en Automatique - Ecole Centrale ParisPacker, BenBenPackerComputer Science Department - Stanford UniversityKoller, DaphneDaphneKollerComputer Science Department [Stanford] - Stanford UniversityMODELING LATENT VARIABLE UNCERTAINTY FOR LOSS-BASED LEARNINGHAL CCSD2012[INFO.INFO-LG] Computer Science [cs]/Machine Learning [cs.LG]Kumar, M. Pawan2013-01-14 13:39:592022-02-03 03:01:402013-01-15 11:08:30enConference papersapplication/pdf1We consider the problem of parameter estimation using weakly supervised datasets, where a training sample consists of the input and a partially specified annotation, which we refer to as the output. The missing information in the annotation is modeled using latent variables. Previous methods overburden a single distribution with two separate tasks: (i) modeling the uncertainty in the latent variables during training; and (ii) making accurate predictions for the output and the latent variables during testing. We propose a novel framework that separates the demands of the two tasks using two distributions: (i) a conditional distribution to model the uncertainty of the latent variables for a given input-output pair; and (ii) a delta distribution to predict the output and the latent variables for a given input. During learning, we encourage agreement between the two distributions by minimizing a loss-based dissimilarity coefficient. Our approach generalizes latent SVM in two important ways: (i) it models the uncertainty over latent variables instead of relying on a pointwise estimate; and (ii) it allows the use of loss functions that depend on latent variables, which greatly increases its applicability. We demonstrate the efficacy of our approach on two challenging problems---object detection and action detection---using publicly available datasets.