Awni Altabaa

prof_pic.jpg

Kline Tower, Office 1117

219 Prospect St

New Haven, CT 06511

Hi! Welcome to my homepage.

My name is Awni. I am a second year PhD student in the Department of Statistics & Data Science at Yale University. I am grateful to be advised by Prof. John Lafferty.

My research interests lie in statistical machine learning, the interface of machine learning and neuroscience, and reinforcement learning theory. My recent projects have explored 1) models of abstraction via relational representation learning; 2) the role of information structures in reinforcement learning; and 3) memory and retrieval mechanisms in machine learning.

selected publications

  1. ML
    Disentangling and Integrating Relational and Sensory Information in Transformer Architectures
    Awni Altabaa, and John Lafferty
    Under Review, 2024
  2. RL
    On the Role of Information Structure in Reinforcement Learning for Partially-Observable Sequential Teams and Games
    Awni Altabaa, and Zhuoran Yang
    Neural Information Processing Systems (NeurIPS), 2024
  3. ML
    Approximation of Relation Functions and Attention Mechanisms
    Awni Altabaa, and John Lafferty
    Under Review, 2024
  4. ML
    Learning Hierarchical Relational Representations through Relational Convolutions
    Awni Altabaa, and John Lafferty
    Transactions on Machine Learning Research (TMLR), 2024
  5. Neuro/ML
    The Relational Bottleneck as an Inductive Bias for Efficient Abstraction
    Taylor W. Webb, Steven M. Frankland, Awni Altabaa, and 6 more authors
    Trends in Cognitive Science (TICS), 2024
  6. ML
    Abstractors and Relational Cross-Attention: An Inductive Bias for Explicit Relational Reasoning in Transformers
    Awni Altabaa, Taylor Webb, Jonathan Cohen, and 1 more author
    International Conference on Learning Representations (ICLR), 2024
  7. RL
    Decentralized Multi-Agent Reinforcement Learning for Continuous-Space Stochastic Games
    Awni Altabaa, Bora Yongacoglu, and Serdar Yüksel
    2023 IEEE American Control Conference (ACC), Mar 2023