- This event has passed.
Seminar: TinyML – the convergence between Machine Learning and the Internet of Things: an end-to-end solution under the Internet of Intelligent Vehicles and Sustainable Cities
June 30 - July 31
Prof. Ivanovitch Silva, Digital Metropolis Institute (IMD,UFRN)
In recent years there has been an increasing discussion on sustainable city planning. In this model, approaches to urban development are based on policies that stimulate citizens’ effective participation and the combination of human, collective, and artificial intelligence to solve everyday problems. Parallel to that, the automotive industry has focused on developing smart vehicles equipped with a variety of sensors. In this way, with the help of an Onboard Diagnostics (OBD-II) device and low-power wide-area network technologies, it is possible to extract and analyze the generated data (edge computing) and apply it in the solution of numerous problems in the context of Sustainable Cities. This scenario leads to a revolution in the way vehicles are used. For example, it is known that the pavement condition of streets and highways is directly related to the driving experience, comfort, and road safety. For this reason, the government must be able to preserve the pavement in a good state. Awareness of anomalies (holes and cracks) in the pavement surface is one of the first steps towards a restoration process, but conducting regular surveys can prove to be either too expensive or unreliable. In the other perspective, data captured by the Mass Air Flow (MAF) sensor (provided by OBD-II device) can be used to measure pollution emitted. It is known that the emission of vehicular gases is one of the significant contributors to air pollution in large urban centers, with pollution indices similar to the industry. These are potential applications only to cite a couple of them. Given this context and the fact that vehicle data can be characterized as a continually evolving data stream, this seminar’s objective is to present the development of a new field of study known as TinyML (Tiny Machine Learning). It is a class of Machine Learning algorithms that can deploy models in low-power devices, as is the case with those found in Vehicles. The solutions will be discussed under the following approach to detect and classify different anomalies in real-time considering aspects and challenges such as instrumentation and embedded hardware, communication protocols, and intelligent algorithms.
Short CV – Prof. Ivanovitch Silva
Ivanovitch Silva received the licentiate, M.Sc., and Ph.D. degrees in Electrical and Computer Engineering from the Federal University of Rio Grande do Norte (UFRN), Natal, Brazil, in 2006, 2008, and 2013. He concluded in 2016 a short course about Big Data & Social Analytics at Massachusetts Institute of Technology (MIT). Since 2013 is professor at Digital Metropolis Institute (IMD,UFRN). He teaches and supervises Ph.D and master students in the Graduate Program of Electrical and Computer Engineering at UFRN. At present, he acts as the coordinator in the Lato Sensu Specialization in Big Data & Analytics at UFRN. His research interests include modeling and scientific data analysis, Internet of Things, Industry 4.0 and Smart Cities.
Available online from 30 June 2021