Volnei A. Pedroni
Federal Technological University of Parana State (UTFPR)
Artificial Intelligence: Why EE students should learn it (internally)
Purpose: Artificial Intelligence (AI) has become a major source of novel, sophisticated solutions for complex engineering problems, from object detection to autonomous vehicles, smart manufacturing to cancer diagnosis, speech detection to text translation, and more. And that includes also the microelectronics field, with AI already present in a number of tasks, including chip design, fabrication, and testing (layout optimization, fabrication processes control, defect detection, etc.). As a consequence, AI has become a tool not only desirable, but fundamental to the whole electrical engineering community. Discussing its adoption as part of the EE curriculum is then highly recommended. The courses should include both machine learning and deep learning (the current tendency is to reserve the “AI” designation for the latter). Both are important because they solve different problems, and indeed they are found together in a number of situations (medical diagnosis, for example). Being a multidisciplinary matter, learning the inner workings of AI in depth is not a trivial endeavor, so courses should be well planed and prepared. On the other hand, the resulting engineers would be qualified to participate in the game, even as future inventor. To call attention to this is the main purpose of this tutorial.
Presentation: The time is very short (75 min) for such an ambitious task, so the selection of content and details was a hard task. In principle, the presentation will consist of four parts. To be sure that participants that are not familiar with AI construction can appreciate the proposal, a (substantial) review will be given in the initial two parts, the first briefly describing the AI groups/subgroups and their respective purposes, and the second reviewing the AI history + construction, from the very beginning until the full control over FFNN (Feedforward Neural Network); these two parts are important to understand and appreciate what follows. In the next two parts, two (surprise) applications, with very interesting characteristics and very much used, are described. To close the presentation, the situation of AI teaching at several top universities will be listed. This, however, should not be the main or only reason for one to embrace or dismiss this initiative.
Volnei A. Pedroni received the BSc degree in Electrical Engineering from the Federal University of Rio Grande do Sul (UFRGS, 1975), and both the MSc (1990) and PhD (1995) degrees in Electrical Engineering from the California Institute of Technology (Caltech), USA. His area of expertise is Microelectronics/Chip Design, having the development of dedicated neural network chips (for AI) been at the core of his PhD work. He was the founder of LME (2010), the microelectronics lab of UTFPR, which allowed UTFPR to be one of the first institutions in Brazil to provide chip fabrication (via MOSIS) to regular undergrad students. The areas of ASICs, FPGAs, and Circuit Design with VHD are of particular interest. He has a number of publications in the field, including two books by MIT Press: Circuit Design with VHDL (a bestseller, in its 3rd edition) and Finite State Machines in Hardware: Theory and Design, with VHDL and SystemVerilog. Prof. Pedroni did collaboration with Caltech (USA), University of Trento (IT), and University of Modena (IT). After retiring from UTFPR in 2017 (but continuing associated for two more years as a Volunteer Professor), he became a regular Invited Professor in the Department of Electrical Engineering of Caltech, a work that lasted until 2025. Recently, Prof. Pedroni has also acted occasionally as a tutor on “Inside AI”. The main courses taught by Prof. Pedroni are the following: At UTFPR: Semiconductor Devices, Electronic Amplifiers, Signals and Systems, Microelectronics (chip design and layout), Statistics, Digital Circuit Design with VHDL. At Caltech: Semiconductor Devices, Analog Circuit Design, Digital Circuit Design with VHDL, Advanced Digital Circuit Design with VHDL. At University of Trento and University of Modena: Digital Circuit Design with VHDL.