Milos Gligoric
University of Texas at Austin, TX·Computer Science
Software testing is the main way developers check that code works, but today's tests are built around whole functions, interactions between functions, or entire systems. Many bugs hide at finer levels, such as single statements inside a function, and slip past these tests.
- Fields
Computer Science
Engineering
- Topics
software testing
software engineering
fine-grained testing
- Grants
- 2 known grants, through Sep 2028
What they work on
- Develop a language and framework for writing fine-grained tests that check individual statements and code fragments inside functions, and evaluate it on open-source projects.
- Automatically generate fine-grained tests from existing code or tests, and adapt them as software evolves to improve regression testing.
- Use fine-grained tests to improve fuzzing and runtime verification, and begin applying them to specific security and performance bugs.
- Build natural language processing models that use execution data to generate and complete unit tests and update them when the code under test changes.
- Create NLP models that migrate tests across programming languages, applied to both general codebases and machine learning applications.
More about the research
Writing and maintaining tests is also slow, tedious work. This research looks for ways to make testing both finer-grained and less labor-intensive, through two collaborative projects: one creates a new kind of test that targets small pieces of code, and the other uses AI language models to help write and update tests.
Why it matters
Software bugs that escape testing cost companies time and money and can cause security and performance failures. Tools that let developers test small code fragments and that use AI to write and maintain tests could help software engineers catch more bugs with less effort, leading to more reliable software in industry and open-source projects.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Benjamin Pierce
University of Pennsylvania, PA·Computer Science
Study how professional developers use property-based testing through HCI studies in industry and real-time observation of developers working with PBT tools.
Yueqi Chen
University of Colorado Boulder, CO·Computer Science
Refactor the Community Earth System Model (CESM) codebase by decoupling its computational and control logic and rewriting the control layer in Python, so that modern security practices can be integrated.
Indrakshi Ray
Colorado State University-Fort Collins, CO·Computer Science
Develop natural language processing models for cross-genre entailment that connect the meaning of privacy laws to technical software specifications, in order to identify privacy vulnerabilities in health data handling.
The school
University of Texas at Austin: 49 labs on PI Pages, $1.1B a year on research, 871 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.