Network Classifiers Based on Social Learning

TL;DR

This paper introduces a social learning-inspired framework for combining classifiers over space and time, enabling continuous improvement during testing with unlabeled data, and demonstrating robustness and consistency.

Contribution

It proposes the Social Machine Learning paradigm that leverages imperfect models for improved, robust classification in streaming data scenarios.

Findings

Improved prediction performance over time with unlabeled data

Robustness against poorly trained classifiers

Theoretical guarantees of consistent learning

Abstract

This work proposes a new way of combining independently trained classifiers over space and time. Combination over space means that the outputs of spatially distributed classifiers are aggregated. Combination over time means that the classifiers respond to streaming data during testing and continue to improve their performance even during this phase. By doing so, the proposed architecture is able to improve prediction performance over time with unlabeled data. Inspired by social learning algorithms, which require prior knowledge of the observations distribution, we propose a Social Machine Learning (SML) paradigm that is able to exploit the imperfect models generated during the learning phase. We show that this strategy results in consistent learning with high probability, and it yields a robust structure against poorly trained classifiers. Simulations with an ensemble of feedforward neural networks are provided to illustrate the theoretical results.