Abstract:
The intelligent edge promises the processing of the data where it is generated, particularly using machine learning and artificial intelligence algorithms. Moving the computation from cloud to edge promises benefits, as latency, costs, and security and privacy risks are reduced, but imposes challenges due to the complexity of the application and the resource constraints found in edge devices. Approximate computing, a low-power, low-energy design paradigm, proposes to reduce resource consumption while lowering the precision of error-tolerant applications, such as those required for the intelligent edge, but keeping the quality-of-results to acceptable levels. This talk will present how approximate computing techniques, at software and hardware level, are enabling the intelligent edge in an efficient and reliable manner.

Bio:
Dr. Jorge Castro-Godínez is an Associate Professor at the School of Electronics Engineering, Costa Rica Institute of Technology (TEC). H received his doctoral degree in Informatics from Karlsruhe Institute of Technology (KIT), Germany, in 2020. He received his Licenciatura (5-year engineering degree) and M.Sc. in Electronics Engineering from Costa Rica Institute of Technology (TEC), in 2011 and 2014, respectively. His primary research interest is in energy-efficient computing across the different abstraction layers, with a particular interest on approximate computing and machine learning on edge computing.