Jeremy Kun

Jianming Tong, Jingtian Dang, Simon Langowski et al.

Zero-knowledge proof (ZKP) provers remain costly because multi-scalar multiplication (MSM) and number-theoretic transforms (NTTs) dominate runtime as they need significant computation. AI ASICs such as TPUs provide massive matrix throughput and SotA energy efficiency. We present MORPH, the first framework that reformulates ZKP kernels to match AI-ASIC execution. We introduce Big-T complexity, a hardware-aware complexity model that exposes heterogeneous bottlenecks and layout-transformation costs ignored by Big-O. Guided by this analysis, (1) at arithmetic level, MORPH develops an MXU-centric extended-RNS lazy reduction that converts high-precision modular arithmetic into dense low-precision GEMMs, eliminating all carry chains, and (2) at dataflow level, MORPH constructs a unified-sharding layout-stationary TPU Pippenger MSM and optimized 3/5-step NTT that avoid on-TPU shuffles to minimize costly memory reorganization. Implemented in JAX, MORPH enables TPUv6e8 to achieve up-to 10x higher throughput on NTT and comparable throughput on MSM than GZKP. Our code: https://github.com/EfficientPPML/MORPH.

Shruthi Gorantala, Jianming Tong, Asra Ali et al.

The deployment of Fully Homomorphic Encryption (FHE) at scale is hindered due to its heavy computational overhead. While specialized hardware accelerators like Google Tensor Processing Units (TPUs) can help, mapping complex cryptographic kernels onto such architectures remains a challenge. Efficient execution requires co-optimization between the systolic array-based Matrix Multiplication Unit (MXU) and Vector Processing Units (VPUs), as well as the orchestration of data movement across the vector register files. Existing compiler stacks often abstract low-level hardware utilization, requiring developers to adopt a manual trial-and-error process that often results in fragmented execution and underutilized resources. To accelerate this development process, we use AlphaEvolve to automate the exploration of hardware-aware cryptographic-kernel optimizations. We frame optimization as an evolutionary search problem, utilizing the closed-loop system provided by AlphaEvolve, that leverages LLM-driven code generation. We use real-world feedback from hardware execution and rigorous correctness testing to guide the evolution process. We evaluate AlphaEvolve optimization on primitives for both the TFHE (Jaxite) and CKKS (CROSS) FHE schemes on Google Cloud TPUv5e, a contemporary TPU architecture. Within 24 hours of automated exploration, AlphaEvolve discovered implementation-level optimizations that improve TFHE bootstrap latency by 2.5x and CKKS rotation and multiplication latency by 1.31x and 1.18x, respectively, relative to human-engineered state of the art. These results demonstrate that AlphaEvolve can be used to enable researchers to navigate the optimization trade-offs between cryptography, compilers, and hardware accelerators.

Benjamin Fish, Jeremy Kun, Ádám D. Lelkes

We study three classical machine learning algorithms in the context of algorithmic fairness: adaptive boosting, support vector machines, and logistic regression. Our goal is to maintain the high accuracy of these learning algorithms while reducing the degree to which they discriminate against individuals because of their membership in a protected group. Our first contribution is a method for achieving fairness by shifting the decision boundary for the protected group. The method is based on the theory of margins for boosting. Our method performs comparably to or outperforms previous algorithms in the fairness literature in terms of accuracy and low discrimination, while simultaneously allowing for a fast and transparent quantification of the trade-off between bias and error. Our second contribution addresses the shortcomings of the bias-error trade-off studied in most of the algorithmic fairness literature. We demonstrate that even hopelessly naive modifications of a biased algorithm, which cannot be reasonably said to be fair, can still achieve low bias and high accuracy. To help to distinguish between these naive algorithms and more sensible algorithms we propose a new measure of fairness, called resilience to random bias (RRB). We demonstrate that RRB distinguishes well between our naive and sensible fairness algorithms. RRB together with bias and accuracy provides a more complete picture of the fairness of an algorithm.

David Burstein, Franklin Kenter, Jeremy Kun et al.

Despite the large effort devoted to cybersecurity research over the last decades, cyber intrusions and attacks are still increasing. With respect to routing networks, route hijacking has highlighted the need to reexamine the existing protocols that govern traffic routing. In particular, our pri- mary question is how the topology of a network affects the susceptibility of a routing protocol to endogenous route misdirection. In this paper we define and analyze an abstract model of traffic interception (i.e. eavesdropping) in distance-vector routing networks. Specifically, we study al- gorithms that measure the potential of groups of dishonest agents to divert traffic through their infrastructure under the constraint that messages must reach their intended destinations. We relate two variants of our model based on the allowed kinds of lies, define strategies for colluding agents, and prove optimality in special cases. In our main theorem we derive a provably optimal monitoring strategy for subsets of agents in which no two are adjacent, and we extend this strategy to the general case. Finally, we use our results to analyze the susceptibility of real and synthetic networks to endogenous traffic interception. In the Autonomous Systems (AS) graph of the United States, we show that compromising only 18 random nodes in the AS graph surprisingly captures 10% of all traffic paths in the network in expectation when a distance-vector routing protocol is in use.

Jeremy Kun, Rajmonda Caceres, Kevin Carter

Learning the right graph representation from noisy, multi-source data has garnered significant interest in recent years. A central tenet of this problem is relational learning. Here the objective is to incorporate the partial information each data source gives us in a way that captures the true underlying relationships. To address this challenge, we present a general, boosting-inspired framework for combining weak evidence of entity associations into a robust similarity metric. Building on previous work, we explore the extent to which different local quality measurements yield graph representations that are suitable for community detection. We present empirical results on a variety of datasets demonstrating the utility of this framework, especially with respect to real datasets where noise and scale present serious challenges. Finally, we prove a convergence theorem in an ideal setting and outline future research into other application domains.

Rajmonda Caceres, Kevin Carter, Jeremy Kun

Learning the right graph representation from noisy, multisource data has garnered significant interest in recent years. A central tenet of this problem is relational learning. Here the objective is to incorporate the partial information each data source gives us in a way that captures the true underlying relationships. To address this challenge, we present a general, boosting-inspired framework for combining weak evidence of entity associations into a robust similarity metric. We explore the extent to which different quality measurements yield graph representations that are suitable for community detection. We then present empirical results on both synthetic and real datasets demonstrating the utility of this framework. Our framework leads to suitable global graph representations from quality measurements local to each edge. Finally, we discuss future extensions and theoretical considerations of learning useful graph representations from weak feedback in general application settings.