|
DANILO PIETRO PAU Technical director, IEEE and ST Fellow, System Research and Applications, STMicroelectronics |
Design of interoperable deeply quantized neural networks for in-sensor computing
Balancing storage and computing capabilities, accuracy, anomaly detection and classifier model design are great challenges when implementing Artificial Neural Networks on ultra-low power devices such as microcontrollers. Deeply Quantized Neural Networks (DQNNs) offer the most promising solution to these requirements, however, current commercial microcontrollers are not yet able to exploit their potentialities. The design of DQNN and custom energy efficient hardware accelerators therefore may represent a viable alternative in terms of energy efficiency, especially with respect to in-sensing neural computing. This results on Hybrid Neural Networks variants developed with experimental deep learning tools, able to achieve high accuracies. Custom ultra-low power hardware circuitry for real-time execution of the Hybrid Neural Network is presented with CMOS technologies and implemented with a Field-programmable gate array, including associated demo.
One year before graduating from the Politecnico di Milano in 1992, Danilo PAU joined STMicroelectronics, where he worked on HDMAC and MPEG2 video memory reduction, video coding, embedded graphics, and computer vision. Today, his work focuses on developing solutions for deep learning tools and applications for resource constrained devices.
Danilo has been an IEEE Fellow since 2019, serving as Industry Ambassador coordinator for IEEE Region 8 South Europe and Member of the Machine Learning, Deep Learning and AI in the CE (MDA) Technical Stream Committee IEEE Consumer Electronics Society (CESoc).
With 81 patents, 106 publications (h-index 24, i10-index 53), 113 MPEG authored documents and 40 invited talks/seminars at various worldwide Universities and Conferences, Danilo's favourite activity remains mentoring undergraduate students, MSc engineers and PhD students from many universities.
Danilo has been an IEEE Fellow since 2019, serving as Industry Ambassador coordinator for IEEE Region 8 South Europe and Member of the Machine Learning, Deep Learning and AI in the CE (MDA) Technical Stream Committee IEEE Consumer Electronics Society (CESoc).
With 81 patents, 106 publications (h-index 24, i10-index 53), 113 MPEG authored documents and 40 invited talks/seminars at various worldwide Universities and Conferences, Danilo's favourite activity remains mentoring undergraduate students, MSc engineers and PhD students from many universities.