ANR JCJC Project: Human-O | 01/2024 - 12/2027
 |
Personalized Optimization with Human Feedback (ANR-23-CE48-0011) |
Abstract
The Human-O project deals with upgrading current optimization algorithms to the needs of cyber-physical and social system, namely information streams and human presence. The aim is to design novel algorithms that learns how optimization problems evolve in time, due to time-varying conditions, and learns human-specific objectives and constraints to be incorporated in the problem itself. The research steps of the project pertain learning the dynamical system underlying the optimality conditions, incorporating human feedback by designing and learning human-specific costs and constraints, and giving humans an active part in the optimization process.
Topics of interests are: convex optimisation, first order algorithms, online learning, kernels, robust optimisation, cyber-physical systems.
People
- Prof. A. Simonetto, (PI)
- Manh Hung Le, postdoc
- Qinyan Zhou, PhD candidate
External experts
- Prof. Emiliano Dall'Anese, Boston University, MA, USA
- Dr. Alessandro Rudi, INRIA, ENS Paris, France
- Prof. Lucian Busoniu, Technical University of Cluj-Napoca, Romania