PhD position in Explainable Combinatorial Optimisation

Updated: almost 2 years ago
Job Type: Temporary
Deadline: 14 Jun 2022

We are seeking a highly motivated candidate to work on the intersection between combinatorial optimisation, AI, and explainability. Modern optimisation methods are invaluable in assisting decision-makers in complex problems with difficult constraints, e.g., scheduling, production-planning, logistics, and transportation. However a glaring problem is that, even though the final solutions provided by the algorithms may be optimised, it may be difficult to understand the reasoning behind the algorithmic decisions. This is particularly a problem in cases where the algorithm is expected to interact with a human decision-maker who needs to justify the decisions to relevant stakeholders.

The goal of the project is to develop novel combinatorial optimisation methods that not only exhibit excellent performance, but also concisely break down the main factors behind automated decision-making in a way that is easy for humans to interpret.

The candidate with join the newly formed XAIT Delft AI Lab alongside three other PhD students. The aim of the lab is to develop novel approaches that focus on the explainable aspect of artificial intelligence for transportation and other civil engineering problems and promote AI-related education. XAIT is led by Panchamy Krishnakumari (Faculty of Civil Engineering and Geosciences: Department of Transport and Planning) and Emir Demirović (Faculty of Engineering, Mathematics and Computer Science: Department of Software and Computer Technology).

The position is for five years. The extra year compared to the usual four-year contract accommodates the 20% additional AI-related education related activities.

The PhD student will be supervised by Dr Emir Demirović and Dr. Gonçalo Correia at TU Delft (The Netherlands).



Similar Positions