Papers
2
Total Citations
24
H-Index
2
About
Mara Graziani is a leading researcher at the intersection of explainable artificial intelligence (XAI) and biomedical engineering. Her work focuses on developing interpretable deep learning models, particularly for healthcare applications. Graziani’s major contributions include advancing concept-based explanations for neural networks, enabling clinicians to understand how AI models arrive at diagnoses from medical images. She has pioneered methods to extract human-interpretable concepts from convolutional neural networks, bridging the gap between black-box AI and clinical trust. Her research on multimodal data integration—combining myoelectric, visual, and gaze tracking signals—has been foundational for improving prosthetic control systems. With over 14 citations for her work on the MeGaNe Pro dataset, which addresses the long-standing challenge of robust robotic hand prostheses, Graziani has demonstrated how explainability techniques can enhance both AI transparency and real-world medical device performance. She is also recognized for developing semi-automatic training methods using gaze tracking and accelerometers to improve object recognition in scene camera data. Her work has been published in top venues like MICCAI and ICML, and she actively contributes to the XAI community through tutorials and open-source tools.
Research Focus
Key Achievements
Top Papers
- 1
- 2