Thao Minh LE

Research Fellow
Applied Artificial Intelligence Institute
Deakin University, Australia

Email: thaoyd2@gmail.com / thao.le@deakin.edu.au

My research interests focus on deep learning and machine learning techniques for visual perception, and vision and language reasoning. These capabilities are the key elements required of the next generation of virtual assistant systems. Real-world applications of these systems include security and safety services, healthcare.


  • [Apr 3, 2024] I have been selected as a DAAD Alnet fellow for the Postdoctoral Networking Tour in AI 04/2024. I will be participating in a virtual networking week (15/4-19/4/2024) and later receiving the DAAD's financial and origanizational support to visit German institutions in person to learn about the German AI research community. Please say "Hi" if you are also attending!
  • [Dec 1, 2023] My grant application on video analysis for early detection of Cerebral Palsy has been successful. I will serve as the Lead Chief Investigator for the two-year project with the Cerebral Palsy Alliance Research Foundation.
  • [Sep 30, 2023] Our paper Dynamic Reasoning for Movie QA: A Character-Centric Approach is accepted by Transactions on Multimedia.
  • [Sep 4, 2023] I am a recipient of the Alfred Deakin Medal for (the most outstanding) Doctoral Thesis in 2021.
  • [Aug 19, 2022] Our paper Guiding Visual Question Answering with Attention Priors is accepted at WACV'23, round 1 (Acceptance rate 22%). Pytorch implementation will be available soon.
  • [Jul 9, 2022] Our paper Video Dialog as Conversation about Objects Living in Space-Time is accepted at ECCV'22. Pytorch implementation is be available on Github.
  • [Jun 6, 2022] Thrilled to receive an academic promotion to Research Fellow at Deakin University.
  • [Mar 30, 2022] I gave a talk on Reasoning Over Vision and Language at FPT Software AI Center's webinar.
  • [Sep 16, 2021] I officially got my PhD from Deakin University.
  • [Aug 6, 2021] Our manuscript Hierarchical Conditional Relation Networks for Multimodal Video Question Answering has been accepted for publication in International Journal of Computer Vision (IJCV).
  • [Jun 29, 2021] Our paper GEFA: Early Fusion Approach in Drug-Target Affinity Prediction is accepted to the IEEE/ACM Transactions on Computational Biology and Bioinformatics.
  • [May 10, 2021] Our tutorial From Deep Learning to Deep Reasoning will be held as part of KDD 2021.
  • [May, 2021] I started working for A2I2@Deakin as a postdoctoral researcher after submitting my doctoral thesis titled Deep Neural Networks for Visual Reasoning on May 10, 2021.
  • [May 1, 2021] Our paper Hierarchical Object-oriented Spatio- Temporal Reasoning for Video Question Answering is accepted at IJCAI'21, acceptance rate 13.9% (587/4204). Code will be available soon!
  • [Apr 11, 2021] Our tutorial Neural Machine Reasoning will be held as part of IJCAI 2021.

  • [Older news]