Beyazit Yalcinkaya

Beyazit Yalcinkaya

I'm a CS PhD student at UC Berkeley, working at the intersection of artificial intelligence and formal methods.

About

I am a 4th-year computer science Ph.D. student at University of California, Berkeley, advised by Sanjit A. Seshia. In 2020, I received my B.Sc. in computer engineering from Middle East Technical University in Ankara, Turkey, where I was advised by Ebru Aydin Gol as an undergraduate research assistant for three years. In the summer of 2018, I was a research intern at the Max Planck Institute for Software Systems in Kaiserslautern, Germany, advised by Björn B. Brandenburg. In 2019, I interned at École Polytechnique Fédérale de Lausanne, where I was advised by George Candea.

Research

I work on artificial intelligence (AI) and formal methods, focusing on AI safety and aligment with Frege’s principle (also called compositionality) in mind. I pursue two major directions: (i) developing compositional formal analysis techniques for the scalable verification of black-box systems, and (ii) designing learning frameworks that leverage the compositional and well-defined semantics of formal models.

Teaching

  • In the Fall 2022 semester, I served as a graduate student instructor for EECS 149/249A: Introduction to Embedded Systems at University of California, Berkeley.
  • In the Fall 2017 semester, I served as an undergraduate teaching assistant for CENG 230: Introduction to C Programming at Middle East Technical University.

Publications

  • Beyazit Yalcinkaya*, Niklas Lauffer*, Marcell Vazquez-Chanlatte*, and Sanjit A. Seshia, “Compositional Automata Embeddings for Goal-Conditioned Reinforcement Learning,” to appear in NeurIPS, 2024. *Equal contribution. (paper, webpage)
  • Beyazit Yalcinkaya*, Niklas Lauffer*, Marcell Vazquez-Chanlatte, and Sanjit A. Seshia, “Automata Conditioned Reinforcement Learning with Experience Replay,” NeurIPS 2023 Workshop on Goal-Conditioned Reinforcement Learning, 2023. *Equal contribution. (paper)
  • Beyazit Yalcinkaya, Hazem Torfah, Daniel J. Fremont, and Sanjit A. Seshia, “Compositional Simulation-Based Analysis of AI-Based Autonomous Systems for Markovian Specifications,” International Conference on Runtime Verification (RV), 2023. (paper)
  • Beyazit Yalcinkaya, Hazem Torfah, Ankush Desai, and Sanjit A. Seshia, “ULGEN: A Runtime Assurance Framework for Programming Safe Cyber-Physical Systems,” IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems (TCAD), 2023. (paper)
  • Niklas Lauffer*, Beyazit Yalcinkaya*, Marcell Vazquez-Chanlatte, Ameesh Shah, and Sanjit A. Seshia, “Learning deterministic finite automata decompositions from examples and demonstrations,” Formal Methods in Computer-Aided Design (FMCAD), 2022. *Equal contribution. (paper)
  • Mert Ergurtuna, Beyazit Yalcinkaya, and Ebru Aydin Gol, “An automated system repair framework with signal temporal logic,” Acta Informatica, 2022. (paper)
  • Beyazit Yalcinkaya and Ebru Aydin Gol, “Clock reduction in timed automata while preserving design parameters,” International Conference on Formal Methods in Software Engineering (FormaliSE), 2019. (paper)
  • Beyazit Yalcinkaya, Mitra Nasri, and Björn B. Brandenburg, “An exact schedulability test for non-preemptive self-suspending real-time tasks,” Design, Automation and Test in Europe Conference (DATE), 2019. (paper)