Titulo Estágio
Evolutionary approaches for Explainable AI
Local do Estágio
DEI-FCTUC
Enquadramento
Machine Learning approaches are becoming increasingly more prominent in all areas of society. Deep Learning models have become widely popular, being commonly used in chat-bots and image generation, as well as in critical scenarios, such as the medical and legal domains, where it is of paramount importance to understand the reasons that motivate each output. However, their black-box nature and typical large sizes hinder their interpretation. This need led to the creation of the research field of Explainable Artificial Intelligence, with some methods having already been proposed to explain such complex models [1].
1 - Das, A., & Rad, P. (2020). Opportunities and challenges in explainable artificial intelligence (xai): A survey. arXiv preprint arXiv:2006.11371.
Objetivo
The main goal of this dissertation is to improve existing explainability approaches with Evolutionary Computation (EC). EC is a family of Artificial Intelligence methods loosely inspired by the principles of evolution by Natural Selection and Mendel’s genetics. These methods have been successfully applied to difficult problems from many domains. In particular, we aim to develop EC methods that improve existing explainability methods or produce explanations of complex machine learning models. An example of a viable option is to evolve interpretable machine learning models that locally explain the original complex ones.
Plano de Trabalhos - Semestre 1
- Literature review.
- Implementation and test of the most promising approaches.
- Writing of the intermediate report.
Plano de Trabalhos - Semestre 2
- Development of evolutionary approaches for local explainability.
- Validation of the proposed methods and comparison with the existing approaches.
- Writing of the dissertation.
- Writing of a scientific article with the main results.
Condições
The work shall be carried out within CISUC’s Bio-Inspired Artificial Intelligence group (bAI), under the supervision of Prof. João Macedo and Prof. Ernesto Costa. Additionally, eligible students may have the opportunity to receive a scholarship (Bolsa de Investigação para Licenciado) following the monthly stipend guidelines set by Fundação para a Ciência e Tecnologia (FCT).
Orientador
João Macedo / Ernesto Costa
jmacedo@dei.uc.pt 📩