A fast Bayesian Artificial Intelligence reasoning engine for modeling and optimization tasks
Thesis type APPLIED RESEARCH, RESEARCH / EXPERIMENTAL
Build a C++ Bayesian Artificial Intelligence reasoning library resorting to all modern programming paradigms including advanced modelling features, paired with an extremal optimization engine.
The Bayesian reasoning is a field of the Artificial Intelligence deputed to the ability of taking decisions using uses the knowledge of uncertain prediction. We aim at developing a C++ library to support such tools in a modern fashion.
The library is going to be composed by two set of tools:
1. The Bayesian core: it provides all programming tools to generate and analyze BNs. The core is going to be build resorting to all available programming paradigm: multi-threading, mixed CPU-GPU computations, etc.
2. The Extremal Optimization engine: build on top of the core, the engine must resort to all analysis capabilities to exploit any kind of multi-objective optimization. The engine will include a meta-programming layer to describe the optimization rules to support the optimization with the maximum flexibility.
3. Extra: if feasible, the final code could be provided with a Python wrapper too.
Learned Outcomes: Advanced C/C++ programming, multi-thread and mixed computation, Bayesian Theory, Optimization algorithms.
Required skills C/C++ programming, basic parallel computing (from Operating Systems lectures).
Deadline 27/07/2023 PROPONI LA TUA CANDIDATURA