Wang photo
Xiaoyang Wang
Lecturer (Assistant Professor) in Artificial Intelligence
Department of Computer Science
Faculty of Environment, Science and Economy
University of Exeter
Contact Info
Innovation Centre, Office K1
Streatham Campus, Exeter, UK, EX4 4RN
CV / Bio / PhD Thesis (中文) / Projects / Teaching / UoE Staff Page
New 2024 Entry
Funding opportunities for students/researchers
China Scholarship Council and University of Exeter PhD Scholarships (Chinese Students)
EPSRC Funded PhD Studentships (September 2024 entry) (Home and International Students)
  • Faculty of Environment, Science and Economy Studentships (Home Students)
  • Self-funded research project opportunities
  • See HERE for more information.
    I've been selected to join the 2023 cohort of EPSRC - Women's Engineering Society(WES) as an ambassador for women in engineering!
    My research focuses on Machine Perception and Intelligence. Recent research efforts and interests are:
  • Reinforcement Learning (RL) and Meta-RL. RL is learning to take action through interactions with the environment to maximize the cumulative reward. It studies intelligent agents, such as computer programs to play games and smart robots. Meta-learning, or learning-to-learn, is to develop broader intelligence. It aims to efficiently learn algorithms that can quickly learn new tasks, which is more close to learning a skill rather than specific solutions. See below the projects for more detail.
  • AI for next-generation communication networks. Communication networks evolve towards virtualization, openness and intelligence, with emerging techniques such as Open Radio Access Network (O-RAN). Transitioning from vendor-based to open-source solutions, O-RAN is actively embracing the technological revolution brought by AI. I study AI-based mobility analysis, decision making and resource management for future communication networks.
  • Computer vision, sensor fusion and generative models. I have worked on several computer vision problems including visual and remote sensing target detection, tracking and trajectory prediction in urban areas. I also work on sensor fusion and generative models.
  • Other research interests include Explainable AI, e.g., neural-symbolic AI and human-AI interaction. I'm also working on AI for health.
    I'm always looking for great students that are interested in the above topics and the broader area of AI. Please feel free to get in touch!
    Generating Solutions: Exploring Conditional Generative Models for Sequential Decision Making (October 06, 2023)

    Advanced Topics in Deep Learning (Joint talk with Jonathan Thomas, February 08, 2022)

    Deep Reinforcement Learning for Future Open RAN (September 15, 2021)
    NG-CDI: Spotlight on the Future of Networks

    Future Open RAN – Intelligence and Challenges (April 01, 2021)
    SCEEM Research Conference, University of Bristol
    See HERE.
    Playing with Alchemy: A Benchmark and Evaluation for Meta-RL
    Funded by the EPSRC Doctoral Training Partnership (DTP) Vacation Internships scheme, University of Bristol, this project focuses on meta-reinforcement learning (Meta-RL), one of the key enablers of cross-domain and cross-task intelligence. We investigate the effectiveness, efficiency and stability of state-of-the-art Meta-RL methods, including the model-agnostic meta-learning approach (MAML) (Finn, Chelsea, 2017), in the Alchemy, DeepMind environment. We also explore the fast adaptation ability of reinforcement learning models from the training domain to unseen tasks, through customized environments. A short project report is available HERE.
    An EPSRC Prosperity Partnership project (2017-2023) working on creating an agile, resilient network capable of meeting the future needs of our rapidly changing society. This project is a collaboration between University of Bristol, University of Cambridge, Lancaster University, University of Surrey and BT. Working with University of Bristol, our team focuses on deep reinforcement learning for future Open RAN (see a video), digital twins (see a Tech talk from Prof. Robert Piechocki) and dynamic environments (see a video).
    Trajectory Prediction in Shared Spaces using Social force models
    Predicting human trajectory in complex scenes is fundamentally challenging due to the changing dynamics, agents' preferences and navigation styles. Social force-based models have long been proposed to replicate agent behaviours (e.g., pedestrians and cyclists) in continuous time and space. We study the performance of SOTA social force-based models for navigation prediction using the large-scale Stanford Drone Dataset. We particularly focus on shared spaces - spatial zones designed to be shared by different classes of agents, through interactive and non-interactive behaviours.
    Remote Sensing Target Detection and Tracking
    My PhD project is about target detection and tracking in remote sensing images and videos using optimization methods. Currently, my interests are in using generative models and new architectures such as Transformers to perform target detection and tracking. I'm working with Dr Yuhan Liu on this project.
    Please refer to Google Scholar for my full list of publications.
    Code available for the following:
    Infrared small target detection via nonnegativity-constrained variational mode decomposition
    Xiaoyang Wang, Zhenming Peng, Ping Zhang, and Yanmin He.
    IEEE Geoscience and Remote Sensing Letters
    Infrared dim and small target detection based on stable multisubspace learning in heterogeneous scenes
    Xiaoyang Wang, Zhenming Peng, Dehui Kong, and Yanmin He.
    IEEE Transactions on Geoscience and Remote Sensing
    Infrared dim target detection based on total variation regularization and principal component pursuit
    Xiaoyang Wang, Zhenming Peng, Dehui Kong, Ping Zhang, and Yanmin He.
    Image and Vision Computing
    Infrared small dim target detection based on local contrast combined with region saliency
    Xiaoyang Wang, Peng Zhenming, Zhang Ping, et al.
    High Power Laser and Particle Beams
  • This term (Spring 2023) I am co-teaching ECM1414: Data Structures and Algorithms with Dr Zeliang Wang.
  • Before moving to Exeter, I partially taught EMAT31530: Introduction to Artificial Intelligence (Fall 2020 and Spring 2021) at the University of Bristol. Teaching materials on Reinforcement Learning are available upon request.
    SCEEM Post-Pandemic Research Restart Events Scheme (SPPRRES) Fund, University of Bristol, 2022.
    Grant value: £1,300
    Engineering Faculty Post-Doctoral Research Prize, University of Bristol, 2021
    Grant value: £5,000
    Bristol Plus Award, University of Bristol, 2018
    Huawei PhD Fellowship, 2017-2018
    Professional Services:
  • Guest Editor for Special Issue "Infrared Sensing and Target Detection" with Dr Yuhan Liu.
  • Reviewer for Conferences: CVPR, BMVC, ECCV, ACCV, ICCV, WACV, etc
  • Reviewer for Journals: IEEE Transactions on Image Processing, IEEE Transactions on Geoscience and Remote Sensing, Scientific Reports, IEEE Signal Processing Letters etc.
  • Technical Program Committee member: VTC 2022 Fall
  • Appointments:
  • Dec 2022 - Present, Honorary Research Associate, Department of Electrical and Electronic Engineering, University of Bristol
  • Nov 2018 - Nov 2022, Research Associate, Department of Electrical and Electronic Engineering, University of Bristol
  • Life outside of work...

    (Watching) Road cycling occasionally ride a bike myself and hiking.
    I contribute to a figure skater fan page of Boyang Jin (nickname "Tiantian"). Follow us on Bilibili and Weibo.

    Inspiration from and thanks to Dr Kayvon Fatahalian's homepage and Sebastin Santy's homepage