Visual Concept Networks: A Graph-Based Approach to Detecting Anomalous Data in Deep Neural Networks

TL;DR

This paper presents a graph-based method that converts images into networks of visual concepts to improve the detection of anomalous and out-of-distribution data in deep neural networks, addressing limitations of current benchmarks.

Contribution

It introduces a novel graph-structured approach utilizing topological features for robust OOD detection, applicable to complex real-world scenarios.

Findings

Effective detection of far-OOD and near-OOD data.

Robust performance across diverse tasks and large vocabularies.

Enhanced DNN resilience to anomalous data.

Abstract

Deep neural networks (DNNs), while increasingly deployed in many applications, struggle with robustness against anomalous and out-of-distribution (OOD) data. Current OOD benchmarks often oversimplify, focusing on single-object tasks and not fully representing complex real-world anomalies. This paper introduces a new, straightforward method employing graph structures and topological features to effectively detect both far-OOD and near-OOD data. We convert images into networks of interconnected human understandable features or visual concepts. Through extensive testing on two novel tasks, including ablation studies with large vocabularies and diverse tasks, we demonstrate the method's effectiveness. This approach enhances DNN resilience to OOD data and promises improved performance in various applications.