The Adaptive Intelligence Lab

team-24-fall.jpeg
University of Southern Denmark
Department of Mathematics and Computer Science (IMADA)
Odense, Denmark

Founded in 2022, we are a brand new lab that exercises basic machine learning research on online, adaptive, and continual reinforcement learning. Our mission is to lift the borders between the training time and the usage time of machine learning algorithms. This way we believe to enable the AI technologies of the future that can perpetually adapt to their users and changing environment conditions throughout their whole life cycle. We define the set of open research questions that hinder real-world use of life-long trainable machine learning artefacts as the adaptive intelligence problem and work hard on solving it.

News

Nov 15, 2024 We have a vacant PhD position:
PhD Position in Reinforcement Learning
Oct 3, 2024 Gulcin Baykal Can starts at our lab as a postdoctoral researcher as of January 2025. Welcome Gulcin!
Oct 3, 2024 Melih will serve as an area chair at AISTATS 2025.

Selected Publications

  1. NeurIPS
    Deterministic Uncertainty Propagation for Improved Model-Based Offline Reinforcement Learning
    In Neural Information Processing Systems 2024
  2. NeurIPS
    Improved Algorithms for Stochastic Linear Bandits Using Tail Bounds for Martingale Mixtures
    H. Flynn, D. Reeb, M. Kandemir, and 1 more author
    In Neural Information Processing Systems 2023
  3. NeurIPS
    Learning Interacting Dynamical Systems with Latent Gaussian Process ODEs
    C. Yildiz, M. Kandemir, and B. Rakitsch
    In Neural Information Processing Systems 2022
  4. ICLR
    Evidential Turing Processes
    M. KandemirA. AkgülM. Haussmann, and 1 more author
    In International Conference on Learning Representations 2022
  5. T-PAMI
    A Deterministic Approximation to Neural SDEs
    A. Look, M. Kandemir, B. Rakitsch, and 1 more author
    IEEE Transactions on Pattern Analysis and Machine Intelligence 2022