PersonalHomeBench: Evaluating Agents in Personalized Smart Homes

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bharadwaj, Manasa, Liu, Yolanda, Yang, InJung, Kim, Sungil, Verma, Nikhil, Kim, KoKeun, Ferreira, Kevin, Kim, YoungJoon
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914564003594240
author Bharadwaj, Manasa
Liu, Yolanda
Yang, InJung
Kim, Sungil
Verma, Nikhil
Kim, KoKeun
Ferreira, Kevin
Kim, YoungJoon
author_facet Bharadwaj, Manasa
Liu, Yolanda
Yang, InJung
Kim, Sungil
Verma, Nikhil
Kim, KoKeun
Ferreira, Kevin
Kim, YoungJoon
contents Agentic AI systems are rapidly advancing toward real-world applications, yet their readiness in complex and personalized environments remains insufficiently characterized. To address this gap, we introduce PersonalHomeBench, a benchmark for evaluating foundation models as agentic assistants in personalized smart home environments. The benchmark is constructed through an iterative process that progressively builds rich household states, which are then used to generate personalized, context-dependent tasks. To support realistic agent-environment interaction, we provide PersonalHomeTools, a comprehensive toolbox enabling household information retrieval, appliance control, and situational understanding. PersonalHomeBench evaluates both reactive and proactive agentic abilities under unimodal and multimodal observations. Thorough experimentation reveals a systematic performance reduction as task complexity increases, with pronounced failures in counterfactual reasoning and under partial observability, where effective tool-based information gathering is required. These results position PersonalHomeBench as a rigorous evaluation platform for analyzing the robustness and limitations of personalized agentic reasoning and planning.
format Preprint
id arxiv_https___arxiv_org_abs_2604_16813
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PersonalHomeBench: Evaluating Agents in Personalized Smart Homes
Bharadwaj, Manasa
Liu, Yolanda
Yang, InJung
Kim, Sungil
Verma, Nikhil
Kim, KoKeun
Ferreira, Kevin
Kim, YoungJoon
Artificial Intelligence
Computation and Language
Databases
Agentic AI systems are rapidly advancing toward real-world applications, yet their readiness in complex and personalized environments remains insufficiently characterized. To address this gap, we introduce PersonalHomeBench, a benchmark for evaluating foundation models as agentic assistants in personalized smart home environments. The benchmark is constructed through an iterative process that progressively builds rich household states, which are then used to generate personalized, context-dependent tasks. To support realistic agent-environment interaction, we provide PersonalHomeTools, a comprehensive toolbox enabling household information retrieval, appliance control, and situational understanding. PersonalHomeBench evaluates both reactive and proactive agentic abilities under unimodal and multimodal observations. Thorough experimentation reveals a systematic performance reduction as task complexity increases, with pronounced failures in counterfactual reasoning and under partial observability, where effective tool-based information gathering is required. These results position PersonalHomeBench as a rigorous evaluation platform for analyzing the robustness and limitations of personalized agentic reasoning and planning.
title PersonalHomeBench: Evaluating Agents in Personalized Smart Homes
topic Artificial Intelligence
Computation and Language
Databases
url https://arxiv.org/abs/2604.16813