Hi! I'm Themis! I am a third-year PhD student at Boston University, jointly advised by Krzysztof Onak and Venkatesh Saligrama. I did my undergrad at Dartmouth College advised by Amit Chakrabarti.

My research focuses on developing efficient and robust algorithms for Foundation Models. I am also interested in problems from different areas of theoretical computer science, including sublinear algorithms, learning theory and complexity.

Publications

Foundation Models
Work on (mostly) Transformer models, efficient inference and interpretability.
  • 1
    Layerwise Dynamics for In-Context Classification in Transformers
    Lutz, Haris, Chandra, Gangrade, Saligrama
    ICML 2026 ★ Spotlight arXiv
  • 2
    Noise Stability of Transformer Models
    Haris, Zhang, Yoshida
    ICLR 2026arXiv
  • 3
    Compression Barriers for Autoregressive Transformers
    Haris, Onak
    COLT 2025arXiv
  • 4
    $k$NN Attention Demystified: A Theoretical Exploration for Scalable Transformers
    Haris
    ICLR 2025 arXiv
Theoretical Computer Science
Work mainly focused on sublinear algorithms for various TCS problems.
Submodular Functions
  • 1
    Efficient Algorithms for Influence Maximization in General Models and Observed Cascades
    Spaeh, Haris, Nguyen, Ene
    IJCAI 2026arXiv
Sublinear Algorithms, Streaming & Property Testing
  • 1
    Testing $k$-submodularity
    Haris, Palit
    RANDOM 2026arXiv
  • 2
    Efficient Algorithms for Adversarially Robust Approximate Nearest Neighbor Search
    Andoni, Haris, Kelman, Onak
    NeurIPS 2025 Workshop arXiv
    • NeurIPS 2025 Workshop: Reliable ML from Unreliable Data
    • WoLA 2026 Poster
  • 3
    Counting Simplices in Hypergraph Streams
    Chakrabarti, Haris
    ESA 2022arXiv
Markov Chains & Spectral Graph Theory
  • 1
    Estimating Hitting Times Locally At Scale
    Haris, Spaeh, Dragazis, Tsourakakis
    NeurIPS 2025arXiv
Differential Privacy
  • 1
    Fast-MWEM: Private Data Release in Sublinear Time
    Haris, Choi, Laksanawisit
    TPDP 2026arXiv
Other
  • 1
    Teaching American Sign Language in Mixed Reality
    Shao, Sniffen, Blanchet, Hillis, Shi, Haris, Liu, Lamberton, Malzkuhn, Quandt, Mahoney, Kraemer, Zhou, Balcom
    IMWUT / UbiCompPaper