Can Reverse Nearest Neighbors Perceive Unknowns?

Payel Sadhukhan

doi:10.1109/access.2019.2963471

Abstract

A novel open set classifier is presented in this work, where the neighborhood of a test instance is determined using the principles of Reverse k-nearest neighbors (RkNN). The RkNN count of an instance can have any non-negative value less or equal to the size of the training set. While dealing with an open dataset, consisting of known and unknown classes, the zero count can provide a possible solution for detecting the unknown class. Positive RkNN count along with the nearest RkNN distance information are used to determine the known class classifications. Experiments are carried out on ten real world datasets, with various openness values on five state-of-the-art open set learners and the proposed scheme. Their performance is measured on three evaluating metrics namely accuracy, average F1 over known and unknown classes, and Known class F1. Empirical results indicate comparable to superior performance delivered by the proposed method over the state-of-the-art approaches on all but one dataset.

Highlights

A conventional classification task aims to assign the instances to any one of the known classes whereas unknown class detection deals with recognition of the instances belonging to unknown classes in addition to the known ones
We present a brief discussion on works that have used principles of reverse nearest neighborhood to achieve some machine learning goals
In this paper, we have presented a novel reverse k-nearest neighbor based classifier

Summary

Introduction

A conventional classification task aims to assign the instances to any one of the known classes whereas unknown class detection deals with recognition of the instances belonging to unknown classes in addition to the known ones. Perception and consequent detection of unknowns pose a serious challenge for the machine, which is designed to operate in a ’closed’ world. We grow and learn in an unknown world with an incrementally growing known subspace whereas our classifiers are trained in a ’closed’ setting of known distributions and classes. It is considered ideal when the training set and the test set have as similar distributions as possible. A classifier is forced to restrict its prediction into the set of training classes. While predicting a test set consisting of seen and unseen class instances, the unseen instances get camouflaged as seen instances and get misclassified

Methods

Results

Conclusion