PhD Course: Binary Analysis with Applications to Machine and Deep Learning
Extracting usable information from executable files is a crucial task in nowadays computing, since it is the basis for a plethora of analyses performed in various fields of computer science and electronics. Analyzing binaries, thus figuring out the true properties of binary programs, however, is not a straightforward task. This is because during the analysis we have to deal with missing symbolic and type information, location-dependent code and data, missing high-level abstraction, mixed code and data and so on.
The course aims at giving a complete overview of the techniques, challenges, tools and applications of binary analysis, introducing fundamentals of both static and dynamic analysis. We will talk about techniques such as disassembling, debugging, binary instrumentation, taint analysis, symbolic execution and we will discuss approaches in security oriented scenarios and in other fields of interest as well.
An important role in the course will be played by machine/deep learning and how these emerging topics can contribute in the context of binary analysis. Specifically, we will talk about (i) various machine/deep learning techniques as versatile approaches to solve binary analysis task, such as similarity detection and vulnerability discovery and (ii) applications of binary analysis as approaches for solving challenging learning tasks, such as authorship attribution and plagiarism detection.
seminar room - 5th floor, cube 42C
22/11/2023 - (09:00-13:00)
23/11/2023 - (09:00-13:00)
24/11/2023 - (09:00-13:00)
Link teams:
12 h - 3 CFU