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

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ganguly, Debargha, Gupta, Debayan, Chaudhary, Vipin
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916412350529536
author Ganguly, Debargha
Gupta, Debayan
Chaudhary, Vipin
author_facet Ganguly, Debargha
Gupta, Debayan
Chaudhary, Vipin
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2409_18235
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Visual Concept Networks: A Graph-Based Approach to Detecting Anomalous Data in Deep Neural Networks
Ganguly, Debargha
Gupta, Debayan
Chaudhary, Vipin
Computer Vision and Pattern Recognition
Machine Learning
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.
title Visual Concept Networks: A Graph-Based Approach to Detecting Anomalous Data in Deep Neural Networks
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2409.18235