Technologies

In addition to a team of experience, AI development requires software technologies and infrastructure. The MAIA laboratory is equipped with state-of-the-art technologies and infrastructure for the application of artificial intelligence.
The lab has a wide range of software technologies and frameworks dedicated to AI. Python and MATLAB programming languages provide the foundation for the implementation and development of machine learning and deep learning algorithms. Libraries such as Pyradiomics, Scikit-learn, and Scipy are used for the extraction and processing of Radiomica. Tools such as TensorFlow, PyTorch, and Keras, and dedicated packages such as cuDNN, facilitate the creation and training of neural networks and sophisticated AI models.
The laboratory’s infrastructure is designed to meet the intensive computing needs required in the field of artificial intelligence. In particular, three stations are equipped with NVIDIA RTX GPUs (48 RT Cores, 384 Tensor Cores, and 3072 CUDA cores with 16 BG GDDR6 memory) to train complex deep learning models, perform simulations and data analysis efficiently and quickly. A collaboration with CINECA is also being activated for the employment of an additional computing platform.