activity
20242026
collaborators

6 papers

cs.LG2026

Noise Stability of Transformer Models

Themistoklis Haris, Zihan Zhang, Yuichi Yoshida

Understanding simplicity biases in deep learning offers a promising path toward developing reliable AI. A common metric for this, inspired by Boolean function analysis, is average…

cs.LG2026

Fast-MWEM: Private Data Release in Sublinear Time

Themistoklis Haris, Steve Choi, Mutiraj Laksanawisit

The Multiplicative Weights Exponential Mechanism (MWEM) is a fundamental iterative framework for private data analysis, with broad applications such as answering linear queries…

cs.DS2026

Efficient Algorithms for Adversarially Robust Approximate Nearest Neighbor Search

Alexandr Andoni, Themistoklis Haris, Esty Kelman +1

We study the Approximate Nearest Neighbor (ANN) problem under a powerful adaptive adversary that controls both the dataset and a sequence of queries. Primarily, for the high-di…

cs.DS2025

Estimating Hitting Times Locally At Scale

Themistoklis Haris, Fabian Spaeh, Spyros Dragazis +1

Hitting times provide a fundamental measure of distance in random processes, quantifying the expected number of steps for a random walk starting at node to reach node . They…

cs.DS2025

Compression Barriers for Autoregressive Transformers

Themistoklis Haris, Krzysztof Onak

A key limitation of autoregressive Transformers is the large memory needed at inference-time to cache all previous key-value (KV) embeddings. Prior works address this by compressin…

cs.LG2024

NN Attention Demystified: A Theoretical Exploration for Scalable Transformers

Themistoklis Haris

Despite their power, Transformers face challenges with long sequences due to the quadratic complexity of self-attention. To address this limitation, methods like -Nearest-Neighb…