Episode 21: AI infrastructures

26 April, 2022

#podcast #T2E13

The artificial intelligence revolution would not have been possible without the adaptation of libraries for computation on graphics cards, also known as GPUs. GPUs made it possible to reduce training time by several orders of magnitude, allowing much more complex and accurate models to be created. Today’s AI requires many of these and other technologies such as TPUs, in large data centers or, as in the case of our institute, in several clusters managed by specialized personnel. In this chapter we talk about the infrastructures that enable Artificial Intelligence with three of our administrators: Ignacio Aguilera Martos (an old acquaintance of our podcast), Germán Gonzalez Almagro and Juan Luis Suárez Díaz.

Listen to our episode!

Links of interest:

• https://www.sciencedirect.com/science/article/pii/S016926072100623 
• https://www.sciencedirect.com/science/article/pii/S0026265X21009796
• https://ityneri.com/actioticket/