On Out-of-Distribution Detection for Audio with Deep Nearest Neighbors

TL;DR

This paper introduces a deep k-nearest neighbors approach for out-of-distribution detection in audio models, significantly improving detection accuracy and false positive rates across various audio datasets.

Contribution

It presents a simple, flexible feature-space method using deep nearest neighbors for effective OOD detection in audio, an under-explored area.

Findings

Improves FPR@TPR95 by 17% over prior methods.

Enhances AUROC score by 7% compared to existing techniques.

Effective across diverse audio and speech datasets.

Abstract

Out-of-distribution (OOD) detection is concerned with identifying data points that do not belong to the same distribution as the model's training data. For the safe deployment of predictive models in a real-world environment, it is critical to avoid making confident predictions on OOD inputs as it can lead to potentially dangerous consequences. However, OOD detection largely remains an under-explored area in the audio (and speech) domain. This is despite the fact that audio is a central modality for many tasks, such as speaker diarization, automatic speech recognition, and sound event detection. To address this, we propose to leverage feature-space of the model with deep k-nearest neighbors to detect OOD samples. We show that this simple and flexible method effectively detects OOD inputs across a broad category of audio (and speech) datasets. Specifically, it improves the false positive rate (FPR@TPR95) by 17% and the AUROC score by 7% than other prior techniques.