Melih Catal
|
|
|
||||||||||||||
Short Bio
I completed my undergraduate education at Sabancı University in Turkey and my master's degree at the University of Zurich. I started my doctoral studies at the University of Zurich in September 2024 under the supervision of Prof. Harald Gall. My research interests are software engineering, software security, machine learning, and cybersecurity.
Publications
Available Thesis and Projects
I am happy to offer Bachelor’s and Master’s theses, as well as Master’s projects, on topics related to Trustworthy CodeLLMs. Several possible directions include:
- Fairness and bias evaluation in code generation
- Machine unlearning techniques in CodeLLMs
- Memorization and data leakage issues in CodeLLMs
- Explainability and interpretability of code generation
If you cannot find a suitable project here but the topics interest you, please do not hesitate to contact me.
Ongoing Thesis and Projects
| Title | Type | Keywords | Description |
| Vertical Integration Bias in Code Generation LLMs | Bachelor's Thesis | Bias and Fairness, Trustworthy AI | Description |
| Differential Privacy Impact on Social Bias in Code Generation | Master's Thesis | Bias and Fairness, Trustworthy AI | Description |
| The Role of Code Comments in the Age of LLMs | Master's Thesis | ESE, Trustworthy AI |
|
| Learn to Unlearn: Machine Unlearning in CodeLLMs | Independent Study | Machine Unlearning, Trustworthy AI | |
| AdvSecureNet: A Python Toolkit for Adversarial Machine Learning | Master Project | Adversarial Machine Learning, Software Development, Trustworthy AI | Description |