Hong Jia
Lecturer (Assistant Professor) at University of Auckland
Honorary Senior Fellow at University of Melbourne
hong.jia AT auckland.ac.nz
Google Scholar - GitHub - LinkedIn - Twitter - University Profile
PhD Students & Visiting Scholars:
I am looking for self-motivated PhD students and visiting scholars to join my research group. Scholarships are available. If you are interested in my research, please feel free to reach out with your CV and research interests. Due to a large number of inquiries, I may not be able to respond to all emails. I will reach out only if I think you are a top candidate and a strong fit for my group.
Students at University of Auckland: I am looking for MSc/MAI/MIT students interested in research projects on AI methods for digital health systems, including: (1) GenAI for Digital Health, (2) Edge and Embodied AI, (3) Efficient ML, Reasoning, and Scaling. Please contact me if you're interested in these areas. You will be working with me and senior PhD students and will have the opportunity to publish in top-tier journals and conferences. Please reach out with your transcript, CV and research interests.
Biography
I am a Lecturer (Assistant Professor) at the University of Auckland and an Honorary Senior Fellow at the University of Melbourne. Previously, I was a Research Associate (2021-2024) at the University of Cambridge. I obtained my Ph.D. in Computer Science from the University of New South Wales (UNSW) in 2021.
My research focuses on AI algorithms and machine learning systems for digital health and edge AI systems, with emphasis on mobile and wearable sensing, personalised health monitoring, mental-health and affective-state modelling, on-device and small language models, uncertainty-aware AI, and human-centred responsible AI. I have served as a PC and reviewer for top-tier journals and conferences, including NeurIPS, ICLR, ICML, AAAI, IJCAI, ACM IMWUT/UbiComp, ACM IPSN, ACM SenSys, and more. My research contributions have been recognized in the field with several awards, including ACM MobiSys Rising Star from ACM SIGMOBILE 2023, Best Paper Runner-Up Award at ACM IPSN 2024, and Best Poster Runner-Up Award at ACM MobiCom 2022. More information about my publications can be found in my Google Scholar.
Research Interests
- Small Language Models: Test-time scaling, reasoning and adaptive scaling strategies for small languge models for reliable and practical deployment.
- Machine Learning for Health: Design machine learning algorithms and systems for digital health systems, including mobile and wearable sensing to enable personalised health monitoring, risk prediction, and timely intervention in real-world settings.
- Edge and Embodied AI Systems: On-device, uncertainty-aware, and human-centred AI for trustworthy interaction, decision-making, and autonomy.
Research Highlights
Two papers were accepted at ICML 2026.
Invited to serve as a Workshop Chair for AJCAI 2026.
We have released the Code of HoRD.
One paper titled "A Framework for On-Device Uncertainty-Aware Event Detection on Microcontrollers" was accepted by Pervasive and Mobile Computing (PMCJ).
Honored to be selected as the ACM TOSN Distinguished Reviewer.
Three papers on test-time scaling and test-time adaptation were accepted at ICASSP 2026.
One paper "Spiking Graph Predictive Coding for Reliable OOD Generalization" was accepted to WWW 2026.
Invited to join the editor board of ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT/UbiComp).
We are organizing the 2nd International Workshop on Mobile Computing with Efficient and Interactive Foundation Models (EIFCom) in conjunction with MobiSys 2026 at Cambridge, UK. Please consider submitting your papers!
Our work "Menta: A Small Language Model for On-Device Mental Health Prediction" has been released. Ref: HuggingFace Papers, Checkpoint, and Code.
Two main track papers and two workshop papers on Efficient ML accepted at NeurIPS 2025. Congrats to my master students Jiaheng (main track, poster, first author), Xin (GenAI4Health workshop, poster, first author) and all coauthors!
Our work "HealthSLM-Bench: Benchmarking Small Language Models for Mobile and Wearable Healthcare Monitoring" on exploring the power of SLMs for Mobile and Wearable Healthcare Monitoring has been accepted by NeurIPS GenAI4Health.
Invited to serve as a TPC member for ICLR 2026, WWW 2026, MobiSys 2026, and ACM SenSys 2026.
One paper titled "Beyond Scale: Small Language Models are Comparable to GPT-4 in Mental Health Understanding" has been accepted by ACII LBR Track.
Invited to serve as a TPC member for ACM SenSys 2026.
One paper titled "LiteFat: Lightweight Spatio-Temporal Graph Learning for Real-Time Driver Fatigue Detection" has been accepted by International Conference on Intelligent Robots and Systems (IROS 2025).
I will be joining the University of Auckland as a Lecturer in July 2025!
One paper titled "Token-Level Logits Matter: A Closer Look at Speech Foundation Models for Ambiguous Emotion Recognition" has been accepted at Interspeech 2025.
Chairing the eighth ACM WellComp 2025 workshop, collocated with ACM IMWUT 2025 in Espoo, Finland. Please consider submitting your work!
Serve as a PC member for NeurIPS 2025, ECAI 2025, and Artifacts Evaluation PC member at ACM MobiCom 2025.
One paper towards AI-Powered Leaf Water Content Monitoring was accepted by IEEE Internet of Things Journal.
One paper titled "LightLLM: A Versatile Large Language Model for Predictive Light Sensing" was accepted by ACM SenSys 2025.
One paper titled "Multimodal Large Language Models in Human-centered Health: Practical Insights" was accepted by IEEE Pervasive Computing.
Invited to serve as a TPC member at ACM MobiSys 2025 .
One paper titled "TinyTTA: Efficient Test-time Adaptation via Early-exit Ensembles on Edge Devices" was accepted by NeurIPS 2024.
A grant worth $11,064 from the Google Research Fund was awarded.
Serve as a PC member for Artifacts Evaluation at ACM SenSys 2024 .
One paper titled "Leveraging LLMs to Predict Affective States via Smartphone Sensor Features" is accepted by Mental Health workshops in conjunction with UbiComp 2024.
One paper titled "StatioCL: Contrastive Learning for Time Series via Non-Stationary and Temporal Contrast" is accepted by ACM CIKM 2024.
One paper titled "Exploring Large-Scale Language Models to Evaluate EEG-Based Multimodal Data for Mental Health" is accepted by WellComp 2024 in conjunction with UbiComp 2024.
Invited talk "Mobile HCI via Edge Intelligence" at UNSW.
Two demo papers using personalized on-device LLMs to analyze human behavior and provide recommendation are accepted by UbiComp 2024.
Our paper LiDARSpectra: Indoor Spectral Mapping with Low-cost LiDARs" has been awarded as the Best Paper Award Runner-up (2/93) by IPSN 2024.
Invited to serve as a TPC member in WellComp 2024, in conjunction with UbiComp 2024.
General Chair and Organizer of the 1st Mobile Computing with Efficient and Interactive Foundation Models (EIFCom) workshop in conjunction with MobiCom 2024.
Invited to serve as Guest Editor of 'Recent Advances in Signal Processing for Flexible and Wearable Electronics' at MDPI Electronics
I have joined the University of Melbourne as a Research Fellow, working with Prof. Vassilis Kostakos. Happy to be back in Australia with my family!
One paper "LiDARSpectra: Indoor Spectral Mapping with Low-cost LiDARs" has been conditionally accepted by IPSN 2024.
One paper "UR2M: Uncertainty and Resource-Aware Event Detection on Microcontrollers" has been accepted by PerCom 2024 (Acceptance rate: 23/158=14.5%).
One paper "LifeLearner: Hardware-Aware Meta Continual Learning System for Embedded Computing Platforms" has been accepted by SenSys 2023. .
Reviewer/TPC for NeurIPS, TMC, TheWebConf, EWSN, UbiComp, IPSN etc.
One paper Spectral-Loc: Indoor Localization using Light Spectral Information has been accepted by IMWUT/UbiComp 2023.
Talk at MobiSys 2023 with the topic "Ubiquitous, Secure, and Efficient Mobile Sensing Systems".
I am honored to be selected as the ACM MobiSys Rising Star (2023) .
One paper Subject-Adaptive Loss-Fitting Smart Garment Platform for Human Activity Recognition has been accepted by TOSN 2023.
One paper on indoor localization Spectral-Loc: Indoor Localization using Light Spectral Information has been accepted by IMWUT/UbiComp 2023.
One paper Turning Silver into Gold: Domain Adaptation with Noisy Labels for Wearable Cardio-Respiratory Fitness Prediction on domain adaptation has been accepted in ML4H 2022 (co-located with NeurIPS 2022).
Our poster paper Passive Light Spectral Indoor Localization has been awarded as the Best Poster Runner-up Award (2/36) by MobiCom 2022.
One paper PROS: An Efficient Pattern-Driven Compressive Sensing Framework For Low Power Biopotential-Based Wearables With On-Chip Intelligence has been accepted by MobiCom 2022! ( Acceptance rate: 17.8%).