Martina Bodini

Training fully connected networks with resistive memories: impact of device failures.

Pritish Narayanan, Stefano Ambrogio, Hsinyu Tsai, 2019, Faraday discussions.

Equivalent-accuracy accelerated neural-network training using analogue memory

Pritish Narayanan, Stefano Ambrogio, Irem Boybat, 2018, Nature.

Improved Deep Neural Network Hardware-Accelerators Based on Non-Volatile-Memory: The Local Gains Technique

Yusuf Leblebici, Pritish Narayanan, Stefano Ambrogio, 2017, 2017 IEEE International Conference on Rebooting Computing (ICRC).