Kaouther Messaoud

Assistant Professor at Télécom Paris

prof_pic.jpg

Welcome! I am an Assistant Professor at Télécom Paris, working on research at the intersection of deep learning, generative models, and autonomous systems. My work focuses on leveraging large-scale multimodal data, representation learning, and cross-domain generalization to advance autonomous vehicle intelligence. I am also increasingly exploring video understanding and its applications to dynamic scene modeling.

Before joining Télécom Paris, I was a Postdoctoral Researcher at the VITA Lab, EPFL, working under the supervision of Prof. Alexandre Alahi. I earned my Ph.D. in 2021 from Sorbonne University, France, where I conducted research on Deep Learning for Motion Prediction at the RITS Lab, INRIA Paris.

News

Mar 01, 2025 Our paper, “Generalizable Trajectory Prediction Using Dual-Level Representation Learning And Adaptive Prompting,” has been accepted at CVPR 2025! Looking forward to presenting our work at CVPR!
Oct 03, 2024 Presented UniTraj Poster at ECCV 2024! It was an incredible opportunity to share our work and engage with researchers from around the world.
Apr 01, 2024 Honored to be an invited lecturer for the course Deep Learning for Autonomous Vehicles at EPFL. Delivered sessions explaining RNNs, Transformers, and their applications in computer vision, LLMs and autonomous driving.
Jul 12, 2023 Proud to receive the 2023 George N. Saridis Best Transactions Paper Award from IEEE TIV!
Apr 15, 2022 Joined the VITA Lab at EPFL as a Postdoctoral Researcher under the guidance of Prof. Alexandre Alahi.

Selected Publications

  1. UniTraj.png
    UniTraj: A Unified Framework for Scalable Vehicle Trajectory Prediction
    Lan Feng, Mohammadhossein Bahari, Kaouther Messaoud, and 3 more authors
    In Computer Vision – ECCV, 2024
    arXiv Code Website
  2. PerReg.png
    Towards Generalizable Trajectory Prediction Using Dual-Level Representation Learning And Adaptive Prompting
    Kaouther Messaoud, Matthieu Cord, and Alexandre Alahi
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2025
  3. PEFT.png
    Forecast-PEFT: Parameter-efficient fine-tuning for pre-trained motion forecasting models
    Jifeng Wang, Kaouther Messaoud, Yuejiang Liu, and 2 more authors
    arXiv preprint arXiv:2407.19564, Jun 2024
    arXiv Code
  4. social_transmotion.png
    Social-Transmotion: Promptable Human Trajectory Prediction
    Saeed Saadatnejad, Yang Gao, Kaouther Messaoud, and 1 more author
    In International Conference on Learning Representations (ICLR), Jun 2024
    arXiv Code
  5. backdoors.png
    Manipulating Trajectory Prediction with Backdoors
    Kaouther Messaoud, Kathrin Grosse, Mickaël Chen, and 3 more authors
    arXiv preprint arXiv:2312.13863, Jun 2023
    arXiv
  6. MHA_JAM.png
    Trajectory Prediction for Autonomous Driving based on Multi-Head Attention with Joint Agent-Map Representation
    Kaouther Messaoud, Nachiket Deo, Mohan M. Trivedi, and 1 more author
    In 2021 IEEE Intelligent Vehicles Symposium (IV), Jun 2021
    DOI arXiv