I am a Postdoctoral Fellow at the Hong Kong PolyU, working with Prof. Dan Wang (HKUST AIS) and Prof. Chen Jason Zhang (PolyU COMP). My research focuses on developing scalable and deployable AI systems for energy systems, spanning smart buildings, HVAC, battery systems, and emerging electrified infrastructures.

I am particularly interested in integrating foundation models and agent-based AI into real-world operations. My recent work explores how AI agents interact with physical systems and operational data for prediction, control, and system optimization. I have been actively involved in, and in some cases led, projects with government and industrial partners in Hong Kong and Japan, focusing on deploying AI systems for practical energy applications.

Before my PhD, I worked as a machine learning engineer at Huawei, JD.com and startups.

Brief CV.

[Google Scholar] [LinkedIn] [ORCID]

Email: yang2.deng@connect.polyu.hk

News

  • May-Jun 2026: Returning as a Visiting Fellow at Prof. Ittetsu Taniguchi’s lab (Osaka University, Japan), working on the projects of smart HVAC sysatem and battery in collaboration with Japanese industry partners.
  • Mar 2026: Our paper “ThermoStill: Distilling Time Series Foundation Model into Thermal Dynamics Model for HVAC Model Predictive Control” is accepted by ACM e-Energy 2026 (Winter round, 18 of 127 submissions were accepted). This is the first work that integrates time series foundation models into MPC-based HVAC control optimization.
  • Jan 2026: Three papers on load prediction interpretability, carbon modeling, foundation model-based carbon forecasting are accepted by ACM e-Energy’26 (Fall round), WWW’26.
  • Oct 2025: Delivered an invited talk titled “Empowering AI Scalability in Building Energy Management Systems” at the Hong Kong Computer Society (香港電脳学会) Artificial Intelligence Seminar: AI in Engineering and Construction (link). Many thanks to Prof. Smason Tai’s invitation.
  • Sep 2025: Three papers on building metadata modeling, carbon modeling, HVAC aggregation control optimization are accepted by Knowledge-Based Systems (KBS), NeurIPS 2025 , ACM BuildSys’25.
  • Aug 2025: Our project on AIoT-based building energy control has been shortlisted for the final assessment in the “PolyU International Future Challenge” entrepreneurship contest.
  • Jul-Aug 2025: I was a Visiting Fellow at Osaka University, Japan, collaborating with Prof. Ittetsu Taniguchi and Dr. Dafang Zhao. During the visit, I led a joint research project among PolyU, Osaka University, and Daikin on optimizing HVAC control.
  • May 2025: Three papers on foundation model and data augmentation got accepted: WeatherFM accepted by IJCAI’25, AugPlug+ accepted by ACM TOSN, FM fine-tuning for building analytics accepted by ICML CO-BUILD’25.
  • Dec 2024: I work as a Postdoctoral Fellow (funded by Research Talent Hub of Innovation and Technology Commission, Hong Kong) in the Department of Computing at PolyU.
  • Nov 2024: Best Ph.D. Forum Presentation Award at ACM BuildSys 2024 in Hangzhou, China! (for my presentation: Improving Cyber-Physical Building Energy System via Large-Scale Machine Learning Evaluation).
  • Oct 2024: Our work AugPlug and two poster/demo are accepted by ACM BuildSys 2024, all related to our BaiTest Project.
  • Sep 2024: I passed my PhD defense! And many thanks to my supervisor Prof. Dan WANG!
  • Aug 2024: Best Presentation Award at the 2nd PolyU Research Student Conference (PRSC 2024).
  • Jun 2024: Best Poster Runner Up at ACM e-Energy 2024 in Singapore!

Selected Publications

  1. MetaCloze: A Schema-guided Automated Building Metadata Model Generation System via Information Extraction
    Fang He, Jiaqi Fan, Yang Deng*, et al. (*corresponding author)
    Knowledge-Based Systems
  2. Concept Drift-aware Time-series Generation for Online Building Load Forecasting: An Automated Data Augmentation Paradigm
    Yang Deng, Rui Liang, Jiaqi Fan, et al.
    ACM Transactions on Sensor Networks (TOSN)
  3. AugPlug: An Automated Data Augmentation Model to Enhance Online Building Load Forecasting
    Yang Deng, Rui Liang, Yaohui Liu, Jiaqi Fan, and Dan Wang
    ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (BuildSys), Best paper candidate, 2024. [pdf] [slides]
  4. Decomposition-based Data Augmentation for Time-series Building Load Data
    Yang Deng, Rui Liang, Dan Wang, Ao Li, and Fu Xiao
    ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (BuildSys), 2023. [pdf] [slides]
  5. Behavior testing of load forecasting models using BuildChecks
    Yang Deng, Jiaqi Fan, Hao Jiang, Fang He, Dan Wang, Ao Li, and Fu Xiao
    ACM International Conference on Future Energy Systems (e-Energy), 2022. [pdf] [slides]
  6. Energon: A Data Acquisition System for Portable Building Analytics
    Fang He, Yang Deng, Yanhui Xu, Cheng Xu, Dezhi Hong, and Dan Wang
    ACM International Conference on Future Energy Systems (e-Energy), 2021. [pdf] [slides]