Enregistré dans:
Détails bibliographiques
Auteurs principaux: Panchal, Sunny, Bhattacharyya, Apratim, Berger, Guillaume, Mercier, Antoine, Bohm, Cornelius, Dietrichkeit, Florian, Pourreza, Reza, Li, Xuanlin, Madan, Pulkit, Lee, Mingu, Todorovich, Mark, Bax, Ingo, Memisevic, Roland
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2407.08101
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911593244131328
author Panchal, Sunny
Bhattacharyya, Apratim
Berger, Guillaume
Mercier, Antoine
Bohm, Cornelius
Dietrichkeit, Florian
Pourreza, Reza
Li, Xuanlin
Madan, Pulkit
Lee, Mingu
Todorovich, Mark
Bax, Ingo
Memisevic, Roland
author_facet Panchal, Sunny
Bhattacharyya, Apratim
Berger, Guillaume
Mercier, Antoine
Bohm, Cornelius
Dietrichkeit, Florian
Pourreza, Reza
Li, Xuanlin
Madan, Pulkit
Lee, Mingu
Todorovich, Mark
Bax, Ingo
Memisevic, Roland
contents Vision-language models have shown impressive progress in recent years. However, existing models are largely limited to turn-based interactions, where each turn must be stepped (i.e., prompted) by the user. Open-ended, asynchronous interactions, where an AI model may proactively deliver timely responses or feedback based on the unfolding situation in real-time, are an open challenge. In this work, we present the QEVD benchmark and dataset, which explores human-AI interaction in the challenging, yet controlled, real-world domain of fitness coaching -- a task which intrinsically requires monitoring live user activity and providing immediate feedback. The benchmark requires vision-language models to recognize complex human actions, identify possible mistakes, and provide appropriate feedback in real-time. Our experiments reveal the limitations of existing state-of-the-art vision-language models for such asynchronous situated interactions. Motivated by this, we propose a simple end-to-end streaming baseline that can respond asynchronously to human actions with appropriate feedback at the appropriate time.
format Preprint
id arxiv_https___arxiv_org_abs_2407_08101
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle What to Say and When to Say it: Live Fitness Coaching as a Testbed for Situated Interaction
Panchal, Sunny
Bhattacharyya, Apratim
Berger, Guillaume
Mercier, Antoine
Bohm, Cornelius
Dietrichkeit, Florian
Pourreza, Reza
Li, Xuanlin
Madan, Pulkit
Lee, Mingu
Todorovich, Mark
Bax, Ingo
Memisevic, Roland
Computer Vision and Pattern Recognition
Vision-language models have shown impressive progress in recent years. However, existing models are largely limited to turn-based interactions, where each turn must be stepped (i.e., prompted) by the user. Open-ended, asynchronous interactions, where an AI model may proactively deliver timely responses or feedback based on the unfolding situation in real-time, are an open challenge. In this work, we present the QEVD benchmark and dataset, which explores human-AI interaction in the challenging, yet controlled, real-world domain of fitness coaching -- a task which intrinsically requires monitoring live user activity and providing immediate feedback. The benchmark requires vision-language models to recognize complex human actions, identify possible mistakes, and provide appropriate feedback in real-time. Our experiments reveal the limitations of existing state-of-the-art vision-language models for such asynchronous situated interactions. Motivated by this, we propose a simple end-to-end streaming baseline that can respond asynchronously to human actions with appropriate feedback at the appropriate time.
title What to Say and When to Say it: Live Fitness Coaching as a Testbed for Situated Interaction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.08101