Mar Castell Erill
Mar Castell Erill is a data scientist with a background in physics, applied mathematics and scientific computing. She pursued her studies in Barcelona and Berlin, gaining experience in several internships in numerical simulations and most recently in a student job at Magnosco GmbH, where she is writing her master’s thesis in methods of improving the generalisability, robustness and uncertainty of deep neural network models in the context of dermatoscopic images. She also engages in graphic design and illustration and actively contributes to various social volunteering projects.
Nov 02, 2023
AI's Silent Influence: Balancing the Wonders and Limitations in …
Convolutional neural networks (CNNs) can recognise objects in images – provided their environment does not change. Mar Castell Erill explains the functionality and limitations of CNNs and makes us wonder whether AI applications are not sometimes overestimated.
Published as part of the Berlin Science Week program 2023
Special Input
Mar Castell Erill
KI und Nachhaltigkeit
en