Negative Object Presence Evaluation (NOPE) to Measure Object Hallucination in Vision-Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lovenia, Holy, Dai, Wenliang, Cahyawijaya, Samuel, Ji, Ziwei, Fung, Pascale
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909285455233024
author Lovenia, Holy
Dai, Wenliang
Cahyawijaya, Samuel
Ji, Ziwei
Fung, Pascale
author_facet Lovenia, Holy
Dai, Wenliang
Cahyawijaya, Samuel
Ji, Ziwei
Fung, Pascale
contents Object hallucination poses a significant challenge in vision-language (VL) models, often leading to the generation of nonsensical or unfaithful responses with non-existent objects. However, the absence of a general measurement for evaluating object hallucination in VL models has hindered our understanding and ability to mitigate this issue. In this work, we present NOPE (Negative Object Presence Evaluation), a novel benchmark designed to assess object hallucination in VL models through visual question answering (VQA). We propose a cost-effective and scalable approach utilizing large language models to generate 29.5k synthetic negative pronoun (NegP) data of high quality for NOPE. We extensively investigate the performance of 10 state-of-the-art VL models in discerning the non-existence of objects in visual questions, where the ground truth answers are denoted as NegP (e.g., "none"). Additionally, we evaluate their standard performance on visual questions on 9 other VQA datasets. Through our experiments, we demonstrate that no VL model is immune to the vulnerability of object hallucination, as all models achieve accuracy below 10\% on NegP. Furthermore, we uncover that lexically diverse visual questions, question types with large scopes, and scene-relevant objects capitalize the risk of object hallucination in VL models.
format Preprint
id arxiv_https___arxiv_org_abs_2310_05338
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Negative Object Presence Evaluation (NOPE) to Measure Object Hallucination in Vision-Language Models
Lovenia, Holy
Dai, Wenliang
Cahyawijaya, Samuel
Ji, Ziwei
Fung, Pascale
Computer Vision and Pattern Recognition
Computation and Language
Object hallucination poses a significant challenge in vision-language (VL) models, often leading to the generation of nonsensical or unfaithful responses with non-existent objects. However, the absence of a general measurement for evaluating object hallucination in VL models has hindered our understanding and ability to mitigate this issue. In this work, we present NOPE (Negative Object Presence Evaluation), a novel benchmark designed to assess object hallucination in VL models through visual question answering (VQA). We propose a cost-effective and scalable approach utilizing large language models to generate 29.5k synthetic negative pronoun (NegP) data of high quality for NOPE. We extensively investigate the performance of 10 state-of-the-art VL models in discerning the non-existence of objects in visual questions, where the ground truth answers are denoted as NegP (e.g., "none"). Additionally, we evaluate their standard performance on visual questions on 9 other VQA datasets. Through our experiments, we demonstrate that no VL model is immune to the vulnerability of object hallucination, as all models achieve accuracy below 10\% on NegP. Furthermore, we uncover that lexically diverse visual questions, question types with large scopes, and scene-relevant objects capitalize the risk of object hallucination in VL models.
title Negative Object Presence Evaluation (NOPE) to Measure Object Hallucination in Vision-Language Models
topic Computer Vision and Pattern Recognition
Computation and Language
url https://arxiv.org/abs/2310.05338