• Corpus ID: 10399946

Confidence-Constrained Maximum Entropy Framework for Learning from Multi-Instance Data

  title={Confidence-Constrained Maximum Entropy Framework for Learning from Multi-Instance Data},
  author={Behrouz Behmardi and Forrest Briggs and Xiaoli Z. Fern and Raviv Raich},
Multi-instance data, in which each object (bag) contains a collection of instances, are widespread in machine learning, computer vision, bioinformatics, signal processing, and social sciences. We present a maximum entropy (ME) framework for learning from multi-instance data. In this approach each bag is represented as a distribution using the principle of ME. We introduce the concept of confidence-constrained ME (CME) to simultaneously learn the structure of distribution space and infer each… 

mil-benchmarks: Standardized Evaluation of Deep Multiple-Instance Learning Techniques

A series of multiple-instance learning benchmarks generated from MNIST, Fashion-MNIST, and CIFAR10 are introduced and the Noisy-And method with label noise is evaluated to find mixed results with different datasets.



Regularized joint density estimation for multi-instance learning

A distance metric for measuring similarities at the bag level which captures the statistical properties of each bag is proposed and a convex optimization method is provided to learn the metric and compare the results with distance based multi-instance learning algorithms.

Multi-instance Metric Learning

This paper proposes a framework called Multi-Instance MEtric Learning (MIMEL) to learn an appropriate distance under the multi-instance setting, and designs a nonparametric density-estimation-based weighting scheme to assign higher “weights” to the instances that are more likely to be positive in positive bags.

Multi-instance clustering with applications to multi-instance prediction

The problem of unsupervised multi-instance learning is addressed where a multi- instance clustering algorithm named Bamic is proposed and based on the clustering results, a novel multi- instances prediction algorithm named Bartmip is developed.

On Confidence-Constrained Rank Recovery in Topic Models

A confidence-constrained rank minimization (CRM) is proposed to recover the exact number of topics in topic models with theoretical guarantees on recovery probability and mean squared error of the estimation.

Multi-Task Learning for Analyzing and Sorting Large Databases of Sequential Data

The MTL-learned task-level sharing mechanisms are used to define the affinity matrix in a graph-diffusion sorting framework that accommodates fast variational Bayesian analysis for large-scale inference, and the effectiveness of the framework is demonstrated using a database composed of 2500 digital music pieces.

Solving the Multiple-Instance Problem: A Lazy Learning Approach

This paper investigates the use of lazy learning and Hausdorff distance to approach the multiple-instance problem, and presents two variants of the K-nearest neighbor algorithm, called Bayesian-KNN and Citation- KNN, solving themultiple- instance problem.

EM-DD: An Improved Multiple-Instance Learning Technique

A new multiple-instance (MI) learning technique that combines EM with the diverse density (DD) algorithm, which provides a new framework for MI learning, in which the MI problem is converted to a single-instance setting by using EM to estimate the instance responsible for the label of the bag.

A Framework for Multiple-Instance Learning

A new general framework, called Diverse Density, is described, which is applied to learn a simple description of a person from a series of images containing that person, to a stock selection problem, and to the drug activity prediction problem.

Multi-Instance Kernels

A kernel on multi-instance data that can be shown to separate positive and negative sets under natural assumptions is shown and compares favorably with state of the art multi- instance learning algorithms in an empirical study.

Rank minimization via online learning

This paper presents the first online learning approach for the problem ofRank minimization of matrices over polyhedral sets and presents two online learning algorithms for rank minimization - the first is a multiplicative update method based on a generalized experts framework, while the second is a novel application of the online convex programming framework.