Mini-Grant POB DigiWorld
Subject: Interpretable neural networks in the context of cognitive processes.
Kwota: 20 000 zł
Abstract.
The project is about employing cache memory theory in the human brain to develop an interpretable deep learning model with prototype layers. The project focuses on developing a model that uses up to nine visual features to describe the data class. AI's interpretability is relevant in applying AI solutions due to the GDPR. GDPR gives the users the law to obtain an understandable explanation of the system's decision.