Source author record

Justin DeBenedetto

Justin DeBenedetto appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

math.NT Cryptography and Security Formal Languages and Automata Theory Machine Learning

Catalog footprint

What is connected

4works

4topics

3close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2020arXiv

Representing Unordered Data Using Complex-Weighted Multiset Automata

Unordered, variable-sized inputs arise in many settings across multiple fields. The ability for set- and multiset-oriented neural networks to handle this type of input has been the focus of much work in recent years. We propose to represent multisets using complex-weighted multiset automata and show how the multiset representations of certain existing neural architectures can be viewed as special cases of ours. Namely, (1) we provide a new theoretical and intuitive justification for the Transformer model's representation of positions using sinusoidal functions, and (2) we extend the DeepSets model to use complex numbers, enabling it to outperform the existing model on an extension of one of their tasks.

preprint2016arXiv

Optimizing Secure Statistical Computations with PICCO

Growth in research collaboration has caused an increased need for sharing of data. However, when this data is private, there is also an increased need for maintaining security and privacy. Secure multi-party computation enables any function to be securely evaluated over private data without revealing any unintended data. A number of tools and compilers have been recently developed to support evaluation of various functionalities over private data. PICCO is one of such compilers that transforms a general-purpose user program into its secure distributed implementation. Here we assess performance of common statistical programs using PICCO. Specifically, we focus on chi-squared and standard deviation computations and optimize user programs for them to assess performance that an informed user might expect from securely evaluating these functions using a general-purpose compiler.

preprint2016arXiv

Quadratic forms representing all integers coprime to 3

Following Bhargava and Hanke's celebrated 290-theorem, we prove a universality theorem for all positive-definite integer-valued quadratic forms that represent all positive integers coprime to $3$. In particular, if a positive-definite quadratic form represents all positive integers coprime to $3$ and $\leq 290$, then it represents all positive integers coprime to $3$. We use similar methods to those used by Rouse to prove (assuming GRH) that a positive-definite quadratic form representing every odd integer between $1$ and $451$ represents all positive odd integers.

preprint2012arXiv

A 60,000 digit prime number of the form $x^{2} + x + 41$

Motivated by Euler's observation that the polynomial $x^{2} + x + 41$ takes on prime values for $0 \leq x \leq 39$, we search for large values of $x$ for which $N = x^{2} + x + 41$ is prime. To apply classical primality proving results based on the factorization of $N-1$, we choose $x$ to have the form $g(y)$, chosen so that $g(y)^{2} + g(y) + 40$ is reducible. Our main result is an explicit, 60,000 digit prime number of the form $x^{2} + x + 41$.