Derivation Miner

Abstract

This report describes Derivation Miner, an effort under the DARPA Mining and Understanding of Software Enclaves (MUSE) program. The research team included Kestrel Institute, Kestrel Technology, UT-Austin, and Qadium. The team investigated the question of whether large online repositories of open source computer code could be used to assist program synthesis in such a way that the result has high assurance. The results of the effort show that this is indeed possible. A variety of indexing, analysis, and search tools have been developed and it has been demonstrated, in many cases, that these tools can be used to identify code in the MUSE corpus that has the desired functionality. Formal proof tools have also been developed, including the Automated Program Transformation (APT) and Axe toolkits, which have been demonstrated over many examples, verifying code found in the wild can be used to synthesize proven programs.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 26, 2019
Accession Number
AD1069544

Entities

People

  • Alessandro Coglio
  • Eric Alden Smith
  • Eric Mccarthy
  • Henny Sipma

Tags

Communities of Interest

  • Autonomy
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Computer Programming
  • Computer Programs
  • Computer Science
  • Computers
  • Dimensionality Reduction
  • Domain Specific Programming Languages
  • Feature Extraction
  • Graphical User Interface
  • Information Retrieval
  • Lisp Programming Language
  • Machine Learning
  • Programming Languages
  • Standards
  • Storage
  • Xml

Fields of Study

  • Computer science
  • Engineering

Readers

  • Distributed Systems and Data Platform Development
  • Software Engineering.