Biography

I obtained my Ph.D degree at School of Computer Science and Engineering, Nanyang Technological University, in July 2023. My research interests include AIoT sensing, smart sensing, and Cybe-physical system. Specifically, I integrate the first principles that guided data generation into machine learning algorithms to address the challenges of data scarcity, labeling or the domain shifts in AIoT applications. My supervisor is Prof Tan Rui. You can find more about our researches at NTU IoT Sensing Group website.

I obtained the bachelor’s degree from School of Electrical and Electronic of Engineering with 2nd upper degree in 2015 at Nanyang Technological University. Thereafter, I spent four years at Micron Technology, Singapore prior starting my PhD degree. I really appreciate about this journey, Micron colleagues are very nice and I’ve had unforgetful memories there.

I am actively looking for job opportunities in both academic and industrial markets.

Download my resumé , 中文版简历 .

Interests
  • Artificial Intelligence of Things
  • Smart Sensing
  • Cyber-physical System
  • Phsyics-guided Learning
Education
  • Ph.D, School of Computer Science and Engineering, 2019.7 - 2023.7

    Nayang Technological Universtiy

  • BSc, Schoole of Electrical and Electronic Engineering, 2011.8 - 2015.6

    Nayang Technological Universtiy

Recent Publications

(2022). Indoor Smartphone SLAM with Learned Echoic Location Features. In SenSys'22.

PDF Project Video

(2022). Sardino: Ultra-Fast Dynamic Ensemble for Secure Visual Sensing at Mobile Edge. In EWSN'22.

PDF Project

(2021). Demo Abstract: Infrastructure-Free Smartphone Indoor Localization Using Room Acoustic Responses. In SenSys'21.

PDF Cite Project Video

(2021). ILLOC: In-Hall Localization with Standard LoRaWAN Uplink Frames. Ubicomp'21.

PDF Cite

(2021). PhyAug: Physics-directed data augmentation for deep sensing model transfer in cyber-physical systems. Best Artifact Award Runner-up, In IPSN'21.

PDF Cite Code Project Video

Researches

*
Location Sensing using Inaudible Echolocation
Smartphone indoor lcoation awareness is increasingly demanded by avariety of mobile applications. The existing solutions for accurate smartphoen indoor location sensisng rely on additional devices or preinstalled infrastructures. This project presents an infrastructure-free smartphone indoor localization system using room acoustic repsonse to a chirp emitted by the phone. Our system can achieve sub-meter localization accuracy.
Location Sensing using Inaudible Echolocation
Physics-directed data augmentation for deep sensing model transfer in cyber-physical systems
Domain shift can greatly degrade the performance of deep model. We exploit the first principle governing the domain shift to augment data for retraining the model. In the case studies of speech recognition, with 5-second unlabelled data collected from the target-domain microphones, we can recover the accuracy losses due to microphone characteristic variations by up to 72%.
Physics-directed data augmentation for deep sensing model transfer in cyber-physical systems