About Me

I am a second year PhD student studying computer science and programming languages at the University of Texas at Austin. My current research is focused on using machine learning to improve program synthesis for source-to-source translation tasks. In my free time, I like to cook, hike, and watch movies.

Contact Details

Gates Dell Complex
2317 Speedway
Austin, TX 78712

bmariano at cs.utexas.edu

Education

University of Texas at Austin

PhD in Computer Science Fall 2019 - Present

University of Maryland

Masters of Computer Science Fall 2017 - Spring 2019

University of Maryland

B.S. Degree in Computer Engineering Fall 2013 - Spring 2017

Research

University of Texas at Austin

PhD student with Isil Dillig Fall 2019 - Present

I am currently working on combining program synthesis and machine learning techniques for automating source-to-source translation problems. In my previous work at UT, I have helped develop a refinement type system for idnetifying arithmetic overflows in smart contracts, create a tool for verifying liveness properties of smart contracts, and design a program synthesis tool for succinctly summarizing common loops in smart contracts.

Max Planck Institute

Intern with Maria Christakis Summer 2019

We developed an optimization methodology for tailoring an abstract interpreter to the code under analysis. In particular, I designed key components of the algorithm, assembled and instrumented real-world benchmarks for testing, and performed detailed analysis of empirical results.

University of Maryland

Master's with Jeff Foster 2017 - 2019

I helped develop JSketch, a sketch-based program synthesis tool for Java. In particular, I designed and tested new models of the Java Standard Library for synthesis optimization. One such method is a novel axiomatic programming paradigm where library behaviour is specified through equational rewriting rules.

Work

University of Texas at Austin

Teaching Assistant for Discrete Math with Isil Dillig Fall 2019

This course covers elementary topics in discrete mathematics for computer science, such as mathematical definitions, logical inference, and proof techniques. I help design examinations and homeworks, grade homeworks, teach discussion, and hold office hours.

University of Maryland

Co-Teacher for Functional Pearls with Cameron Moy Spring 2019

Our class exposed students to elegant examples of functional programming. The first half of the course used Haskell to teach important functional programming abstractions (e.g. Monads), while the second half focused on interesting applications and techniques in programming languages. I developed all course content, including lectures and homeworks, jointly with my co-teacher Cameron.

University of Maryland

Teaching Assistant for Advanced Functional Programming with Niki Vazou Fall 2017

Our class used Haskell to teach undergrads standard functional programming concepts like monoids, functors, and monads. I helped design homeworks and lectures, graded homeworks and tests, and conducted office hours with students. I also taught a lecture on Haskell QuickCheck.

Prime Solutions

Software Engineering Intern Summer 2016

Worked on a team of five interns to create the Advanced Persistant Threat Simulator, a software for simulation of malicious network protocols in C and Python.

Prime Solutions

Software Engineering Intern Summer 2015

Worked on a team of four interns to create the Wi-Fi Traffic Analyzer, a tool for automating analysis of local Wi-Fi traffic to determine device-specific information.



Publications




Automatically Tailoring Static Analysis to Custom Usage Scenarios

N. Mansur, B. Mariano, M. Christakis, J. Navas, V. Wustholz (CAV 2021)

SmartPulse: Automated Checking of Temporal Properties in Smart Contracts

J. Stephens, K. Ferles, B. Mariano, S. Lahiri, I. Dillig (S&P 2021)

Demystifying Loops in Smart Contracts

B. Mariano, Y. Chen, Y. Feng, S. Lahiri, and I. Dillig (ASE 2020)

Program Synthesis with Algebraic Library Specifications

B. Mariano, J. Reese, S. Xu, X. Qiu, T. Nguyen, J. Foster, and A. Solar-Lezama (OOPSLA 2019)

SolType: Refinement Types for Solidity

B. Tan, B. Mariano, S. Lahiri, I. Dillig, Y. Feng (in review)

Imperative-to-Functional API Translation via Neurosymbolic Program Synthesis

B. Mariano, Y. Chen, Y. Feng, G. Durrett, I. Dillig (in preparation)

Address and Email

Ben Mariano
Gates Dell Complex 2317 Speedway
Austin, TX 78712

bmariano at cs.utexas.edu