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

TL;DR

This paper introduces standardized benchmarks for evaluating deep multiple-instance learning techniques using datasets like MNIST, Fashion-MNIST, and CIFAR10, and assesses various methods including the Noisy-And approach under different conditions.

Contribution

It provides a comprehensive framework and benchmarks for evaluating multiple-instance learning methods, along with implementations and evaluations of several techniques.

Findings

Evaluated multiple-instance learning techniques on new benchmarks.

Found mixed results for the Noisy-And method with label noise.

Provided open-source implementations and benchmarks for future research.

Abstract

Multiple-instance learning is a subset of weakly supervised learning where labels are applied to sets of instances rather than the instances themselves. Under the standard assumption, a set is positive only there is if at least one instance in the set which is positive. This paper introduces a series of multiple-instance learning benchmarks generated from MNIST, Fashion-MNIST, and CIFAR10. These benchmarks test the standard, presence, absence, and complex assumptions and provide a framework for future benchmarks to be distributed. I implement and evaluate several multiple-instance learning techniques against the benchmarks. Further, I evaluate the Noisy-And method with label noise and find mixed results with different datasets. The models are implemented in TensorFlow 2.4.1 and are available on GitHub. The benchmarks are available from PyPi as mil-benchmarks and on GitHub.