UINav: A Practical Approach to Train On-Device Automation Agents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Wei, Hsu, Fu-Lin, Bishop, Will, Campbell-Ajala, Folawiyo, Lin, Max, Riva, Oriana
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910505152544768
author Li, Wei
Hsu, Fu-Lin
Bishop, Will
Campbell-Ajala, Folawiyo
Lin, Max
Riva, Oriana
author_facet Li, Wei
Hsu, Fu-Lin
Bishop, Will
Campbell-Ajala, Folawiyo
Lin, Max
Riva, Oriana
contents Automation systems that can autonomously drive application user interfaces to complete user tasks are of great benefit, especially when users are situationally or permanently impaired. Prior automation systems do not produce generalizable models while AI-based automation agents work reliably only in simple, hand-crafted applications or incur high computation costs. We propose UINav, a demonstration-based approach to train automation agents that fit mobile devices, yet achieving high success rates with modest numbers of demonstrations. To reduce the demonstration overhead, UINav uses a referee model that provides users with immediate feedback on tasks where the agent fails, and automatically augments human demonstrations to increase diversity in training data. Our evaluation shows that with only 10 demonstrations UINav can achieve 70% accuracy, and that with enough demonstrations it can surpass 90% accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2312_10170
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle UINav: A Practical Approach to Train On-Device Automation Agents
Li, Wei
Hsu, Fu-Lin
Bishop, Will
Campbell-Ajala, Folawiyo
Lin, Max
Riva, Oriana
Human-Computer Interaction
Artificial Intelligence
Automation systems that can autonomously drive application user interfaces to complete user tasks are of great benefit, especially when users are situationally or permanently impaired. Prior automation systems do not produce generalizable models while AI-based automation agents work reliably only in simple, hand-crafted applications or incur high computation costs. We propose UINav, a demonstration-based approach to train automation agents that fit mobile devices, yet achieving high success rates with modest numbers of demonstrations. To reduce the demonstration overhead, UINav uses a referee model that provides users with immediate feedback on tasks where the agent fails, and automatically augments human demonstrations to increase diversity in training data. Our evaluation shows that with only 10 demonstrations UINav can achieve 70% accuracy, and that with enough demonstrations it can surpass 90% accuracy.
title UINav: A Practical Approach to Train On-Device Automation Agents
topic Human-Computer Interaction
Artificial Intelligence
url https://arxiv.org/abs/2312.10170