AVTA Transition: Toolchain for Binary Decompilation and Alignment

Abstract

The AVTA Project has led to several innovations and tools that have improved the accuracy of binary decompilation [S&P 21,DSN 21], t,he ability to correlate transformed programs [OOPSLA 19,CAV 21], and techniques for verifying temporal properties of decompiled bina,ries [APLAS 19]. Despite these innovations in several aspects of the problem, the state-of-the-art lacks the ability to integrate th,ese three distinct areas all together.We now propose a transition plan, wherein these innovations are brought together, with some mo,dest extensions, to form a unified toolchain for analyzing transformed binaries. In particular, we propose three specific research,Tasks: (Task 1) Improving the accuracy of binary lifting, (Task 2) Automatically aligning transformed binaries, and (Task 3) Reducin,g the complexity of decompiled IR. The main outcomes of this project will be a single toolchain capable of taking input binaries, ac,curately decompiling them, and then automatically correlating the binaries to discover syntactic and semantic differences/similariti,es, as well as scientific results and publications covering solutions to the challenges in each task. The proposed work will lead to, tools that can analyze transformed binaries and detect and isolate what changes have been made, allowing users, e.g. within DoD, to, increase their confidence in the safety and security of those programs.

Document Details

Document Type
DoD Grant Award
Publication Date
Sep 03, 2022
Source ID
N000142212643

Entities

People

  • Eric Koskinen

Organizations

  • Office of Naval Research
  • Stevens Institute of Technology
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Astronomy/Astrophysics
  • Computational Linguistics
  • Systems Analysis and Design