Author ORCID Identifier
https://orcid.org/0009-0001-9901-9099
Date of Award
Spring 5-30-2024
Document Type
Thesis (Undergraduate)
Department
Computer Science
First Advisor
Kyungtae Kim
Abstract
Measuring the confidentiality of programs that need to interact with the outside world can prevent leakages and is important to protect against dangerous attacks. However, information propagation is difficult to follow through a large program with implicit information flow, tricky loops, and complicated instructions. Previous works have tackled this problem in several ways but often measure leakage a program has on average rather than the leakage produced by a set of particularly compromising interactions. We introduce new methods that target a specific set of observables revealed throughout execution to cut down on the resources needed for analysis. Our implementation examines Python bytecode and we compare its results with Meta's Pyre, a static taint analysis engine.
Recommended Citation
Utley, John J., "Measuring Confidentiality with Multiple Observables" (2024). Computer Science Senior Theses. 45.
https://digitalcommons.dartmouth.edu/cs_senior_theses/45
