ViQuAE, a Dataset for Knowledge-based Visual Question Answering about Named Entities - Archive ouverte HAL Access content directly
Conference Papers Year :

ViQuAE, a Dataset for Knowledge-based Visual Question Answering about Named Entities

(1, 2) , (3) , (1, 2) , (3) , (3) , (4, 5) , (4, 5)
1
2
3
4
5

Abstract

Whether to retrieve, answer, translate, or reason, multimodality opens up new challenges and perspectives. In this context, we are interested in answering questions about named entities grounded in a visual context using a Knowledge Base (KB). To benchmark this task, called KVQAE (Knowledge-based Visual Question Answering about named Entities), we provide ViQuAE, a dataset of 3.7K questions paired with images. This is the first KVQAE dataset to cover a wide range of entity types (e.g. persons, landmarks, and products). The dataset is annotated using a semi-automatic method. We also propose a KB composed of 1.5M Wikipedia articles paired with images. To set a baseline on the benchmark, we address KVQAE as a two-stage problem: Information Retrieval and Reading Comprehension, with both zero-and few-shot learning methods. The experiments empirically demonstrate the difficulty of the task, especially when questions are not about persons. This work paves the way for better multimodal entity representations and question answering. The dataset, KB, code, and semi-automatic annotation pipeline are freely available at https://github.com/PaulLerner/ViQuAE.
Fichier principal
Vignette du fichier
lerner_sigir_2022_camera.pdf (9.43 Mo) Télécharger le fichier
Origin : Publisher files allowed on an open archive

Dates and versions

hal-03650618 , version 1 (26-04-2022)

Identifiers

Cite

Paul Lerner, Olivier Ferret, Camille Guinaudeau, Hervé Le Borgne, Romaric Besançon, et al.. ViQuAE, a Dataset for Knowledge-based Visual Question Answering about Named Entities. SIGIR '22: The 45th International ACM SIGIR Conference on Research and Development in Information Retrieval, Jul 2022, Madrid, Spain. ⟨10.1145/3477495.3531753⟩. ⟨hal-03650618⟩
164 View
275 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More