Intervention timing
When should a wearable agent step in, stay quiet, or wait for the user's attention?
WearAgent 2026
The 1st International Workshop on Interactive AI for Personal Wearable Agents
Exploring wearable agents that bridge continuous sensing, foundation model reasoning, and proactive human-agent interaction.
Conference
UbiComp/ISWC 2026
Time
Oct 11-12, 2026
Venue
Shanghai, China
Wearable agents are becoming the next interface between bodies, context, and intelligence.
WearAgent 2026 focuses on a shift from passive activity recognition to proactive, agentic assistance, connecting continuous wearable sensor streams with the reasoning capabilities of LLMs and multimodal foundation models.
Situated agency
How can an agent act in real time on a real body under physical, computational, and attentional constraints?
Human-agent boundary
Where should the boundary sit between human agency, confirmation, autonomy, and proactive assistance?
Affiliated Institutions
Call for Papers
We invite contributions that define the next generation of personal wearable agents.
Topics of Interest
- Sensor-in-the-loop agents: architectures for grounding LLMs and foundation models in continuous IMU, audio, vision, and physiological sensor data.
- Proactive interaction design: non-intrusive, context-aware nudges, audio interfaces, and micro-interactions for screenless or limited-attention wearable form factors.
- On-device reasoning: efficient, low-latency agentic reasoning with small language models, edge AI, and resource-constrained wearables.
- Activity and context understanding: moving beyond HAR toward semantic understanding of user intent, cognitive load, and emotional state.
- Human-agent collaboration: studies on user trust, agency, reliance, interruptibility, and the social comfort of always-on wearable assistants.
- Privacy and ethics: privacy-preserving egocentric sensing, consent, bystander privacy, and ethical continuous personal data processing.
- Systems and prototypes: hardware and software platforms for smart glasses, hearables, smart clothing, and other agentic wearable interfaces.
Accepted Papers
Accepted contributions to the first WearAgent workshop at UbiComp/ISWC 2026.
-
Kinesis: A Multi-Agent Wearable System for Contextualized Posture Correction
Jiayue Ni, Jianing Yu, and Lilith Yu
-
BrickBuddy: An Egocentric Wearable Companion Agent for Step-Situated Cultural Discovery During Procedural Assembly
Xuejiao Ma, Ruijia Li, and Bo Jiang
-
When Does Activity Recognition Error Change Agent Behavior? A Policy-Aware Sensor-to-Decision Interface for Wearable Agents
Chuqi Dai, Xinyao Li, Fanbo Meng, and Zhanpeng Jin
-
ShiftMind: Shared Workforce Memory and Agentic Analysis on Smart Glasses
Jose Manuel Davila Mancilla, Eduardus Tjitrahardja, and Yuntao Wang
-
StressPet: A Sensor-in-the-Loop Agent for Proactive, Just-in-Time Stress Management
Xinyao Li, Fanbo Meng, Chuqi Dai, and Zhanpeng Jin
-
Building Reliable LLM Coding Agents for Mobile Health Data Analysis
Qiang Yang, Alexandra Loh, Yuwei Zhang, and Cecilia Mascolo
-
Beyond Sensor Streams: HALO-Agent as a Body-State Representation Layer for Wearable Agents
Wenbo Zhang, Yuan Gao, Dong Huitian, Chenxu Zhu, Chenxu Zhang, Wenkang Zhang, Yang Gao, Jagmohan Chauhan, and Zhanpeng Jin
-
From Scene Understanding to Spatial Intention: A Personal Wearable Agent for Non-Visual Guidance
Yirui Yuan, Chenghang Li, Shurui Zhang, and Yang Gao
-
AutoHabits: A Close-Looped Context-Aware System for Tiny Habits on Commercial Watches and Phones
Yingke Ding, Zeyu Wang, Hongjing Piao, Leyi Zhang, William Hu, Ziyu Tang, Wanying Mo, Yuanchun Shi, and Yuntao Wang
-
A Mobile–Ambient Sensing System for Cross-Context Activity Understanding
Shukai Ma and Yubin Lan
-
RASE: Reliability-Aware Sustained Evidence for Ring-Based WearAgent Vaping Feedback
Yi-Shan Wu, Eunhee Park, and Yincheng Jin
-
GlassClaw: A Smart-Glasses Agent That Learns When to Help
Wangrui Qian
-
HandContextLLM: Task-Aware Evidence Selection for Continuous Hand-Task Understanding and Forecasting
Chiyue Wang, Dong She, Yang Gao, and Zhanpeng Jin
-
Edge Interactive Agent for Type 1 Diabetes Self-Management with Continuous Glucose Monitoring
Joshua Kramer and Guangzhi Tang
Organizing committee
A cross-disciplinary team spanning wearable computing, ubiquitous sensing, HCI, mobile systems, and agentic AI.
Zhanpeng Jin
South China University of Technology
Homepage
Yang Gao
South China University of Technology
Homepage
Uichin Lee
KAIST
Homepage
Sicong Liu
Northwestern Polytechnical University
Homepage
Dong Ma
University of Cambridge
Homepage
Junyi Zhu
University of Michigan
Homepage
Qian Zhang
Shanghai Jiao Tong University
Homepage
Jagmohan Chauhan
University College London
Homepage
Huatao Xu
ByteDance
HomepageSchedule
Important dates and a half-day program designed around interaction, debate, and community building.
-
Jul 2Jul 18Submission deadlineJuly 2, 2026, 23:59 AOE.July 18, 2026, 23:59 AOE. -
Jul 25Jul 31Notification of acceptanceJuly 25, 2026, 23:59 AOE.July 31, 2026, 23:59 AOE. -
Jul 30Aug 7Camera-ready deadlineJuly 30, 2026, 23:59 AOE.August 7, 2026, 23:59 AOE. -
Oct 11/12
Workshop date
October 11/12, 2026, during UbiComp/ISWC 2026 in Shanghai.
All deadlines are 23:59 AOE.
09:00-09:10
Opening Remarks
Welcome, workshop goals, and framing of the three core research shifts.
09:10-09:40
Keynote
Invited speaker: Prof. Fan Wu, Shanghai Jiao Tong University, confirmation pending.
09:40-10:20
Lightning Talks
Five-minute presentations of accepted papers.
10:20-11:20
Interactive Activity
Agent Envisioning: groups design proactive wearable-agent interaction scenarios.
11:20-12:20
Debate Panel
Proactive vs. Reactive: agency, interruptibility, and the boundaries of always-on sensing.
12:20-12:30
Closing Remarks
Synthesis, awards, and next steps for community building.