Multiple-Instance Learning with Structured Bag Models
نویسندگان
چکیده
Traditional approaches to Multiple-Instance Learning (MIL) operate under the assumption that the instances of a bag are generated independently, and therefore typically learn an instance-level classifier which does not take into account possible dependencies between instances. This assumption is particularly inappropriate in visual data, where spatial dependencies are the norm. We introduce here techniques for incorporating MIL constraints into Conditional Random Field models, thus providing a set of tools for constructing structured bag models, in which spatial (or other) dependencies are represented. Further, we show how Deterministic Annealing, which has proved a successful method for training non-structured MIL models, can also form the basis of training models with structured bags. Results are given on various segmentation tasks.
منابع مشابه
Multiple Instance Learning for bags with Ordered instances
Multiple Instance Learning (MIL) algorithms are designed for problems where labels are available for groups of instances, commonly referred to as bags. In this paper, we consider a new MIL problem setting where instances in a bag are not exchangeable, and a bijection exists between every pair of bags. We propose a neural network based MIL algorithm (MILOrd) that leverages the existence of such ...
متن کاملInstance Label Prediction by Dirichlet Process Multiple Instance Learning
We propose a generative Bayesian model that predicts instance labels from weak (bag-level) supervision. We solve this problem by simultaneously modeling class distributions by Gaussian mixture models and inferring the class labels of positive bag instances that satisfy the multiple instance constraints. We employ Dirichlet process priors on mixture weights to automate model selection, and effic...
متن کاملSalience Assignment for Multiple-Instance Regression
We present a Multiple-Instance Learning (MIL) algorithm for determining the salience of each item in each bag with respect to the bag’s real-valued label. We use an alternating-projections constrained optimization approach to simultaneously learn a regression model and estimate all salience values. We evaluate this algorithm on a significant real-world problem, crop yield modeling, and demonstr...
متن کاملReview of Multi-Instance Learning and Its applications
Multiple Instance Learning (MIL) is proposed as a variation of supervised learning for problems with incomplete knowledge about labels of training examples. In supervised learning, every training instance is assigned with a discrete or real-valued label. In comparison, in MIL the labels are only assigned to bags of instances. In the binary case, a bag is labeled positive if at least one instanc...
متن کاملRevisiting multiple instance neural networks
Recently neural networks and multiple instance learning are both attractive topics in Artificial Intelligence related research fields. Deep neural networks have achieved great success in supervised learning problems, and multiple instance learning as a typical weakly-supervised learning method is effective for many applications in computer vision, biometrics, nature language processing, etc. In...
متن کامل