Pavel Czempin
Student Researcher | Artificial Intelligence
Technical University of Munich
UC Berkeley
Biography
Pavel Czempin is a master’s student at the Technical University of Munich and a former visiting researcher at the Center for Human-Compatible AI at UC Berkeley. His research interests include reinforcement learning, robust AI systems and AI safety.
Interests
- Artificial Intelligence
- AI Robustness
- Deep RL
- AI Safety
Education
MSc in Computer Science (Artificial Intelligence)
Technical University of Munich
BSc in Engineering Science
Technical University of Munich
Research Experience
Visiting Student Researcher
- In Collaboration with Adam Gleave, Daniel Filan
- Research on interpretability for learned reward models.
Research Intern
- Supervised by Adam Gleave
- Research on defenses against adversarial reinforcement learning policies.
Research Assistant
- Supervised by Nicolas Müller
- Implemented and compared audio deepfake detection architectures
Publications
(2022).
Does Audio Deepfake Detection Generalize?.
In Proc. Interspeech 2022.
(2022).
Reducing Exploitability with Population Based Training.
In AdvML Frontiers at ICML.
(2021).
Speech is Silver, Silence is Golden: What do ASVspoof-trained Models Really Learn?.
In ASVspoof2021.