Topic: learning theory

Develop a Python library for evaluating theoretical generalization bounds in deep learning within the SIERRA team and ERC DYNASTY project. The role connects learning theory with practical neural-network evaluation through modular scientific software, empirical experiments and publications. The two-year appointment has a planned start of 1 December 2026. Flexible working and teleworking are possible.

Join Tom Gur’s group to study classical and quantum complexity. Research directions include sublinear algorithms, property testing, cryptography, proof systems, locally testable codes, learning theory, harmonic analysis and additive combinatorics. The one-year project connects theoretical computer science with mathematical and quantum methods.

Katrina Ligett explores why learning theory can make pessimistic predictions even when practical systems succeed. Instead of analysing only worst-case algorithms and data, she proposes reformulating questions around properties that explain algorithmic success and instance-specific performance. Examples in generalisation and privacy illustrate how these perspectives could narrow the gap between theory and empirical practice and create new mathematical questions.