Ms Barikisu is a doctoral researcher in Electrical and Computer Engineering at FEUP, with a focus on classical machine learning applications for deployment on embedded devices. Her research centers on understanding the critical tradeoffs between model accuracy and computational efficiency , identifying models that achieve good performance while remaining lightweight enough for real-world deployment on resource-constrained devices.
Her motivation stems from realizing that not all problems require advanced neural networks. Classical approaches often provide efficient solutions when properly applied. She has researched how training dataset size impacts model performance and computation some models reach good performance with less data, while others demand large datasets. Similarly, some algorithms scale independently of training data size, while others don't
Beyond research, Barikisu is passionate about advocating for ethical AI adoption in this generation of generative AI. She is particularly concerned about risks like digital dementia in younger generations, where over-reliance on AI might erode our capacity for critical reasoning. She believes raising awareness and encouraging thoughtful, responsible use of AI technology is essential.
