A Ph.D. scholarship will be opening soon in Embedded Systems and the Internet of Things at the Department of Information Engineering and Computer Science, University of Trento, Italy.

Research Topic

The objective of this Ph.D. is to design and develop techniques to perform efficient machine learning inference and training on resource-constrained embedded devices. First, we will explore and devise new resource-aware training, model generation, and compression techniques to fit machine-learning models into memory-constrained edge devices. Second, we will build computationally light execution techniques to run the compressed models effectively and in an energy-efficient manner on several low-power microcontroller-based platforms. Third, we will contribute to the benchmarking of tiny machine learning systems' performances in an architecture-neutral, representative, and reproducible manner.

We will also target TinyML on the new generation resource-constrained IoT edge devices, which can even operate without batteries by using ambient energy only, e.g., solar power or radiofrequency harvesting. These devices can store only a small amount of energy in their capacitors and do not include batteries. A big challenge is to keep them sensing, computing, and communicating efficiently under small energy budgets.

The proposed research will be on the intersection of the following topics:

• Energy Efficient, and Low-Power Computing
• Machine Learning on Microcontrollers (Tiny Machine Learning)
• Software for Embedded Systems
• Architectural Support

Mandatory Requirements

• A master's degree (or equivalent) in computer science and engineering, electronics, and electrical engineering.
• English level certification (TOEFL or others). The language of our research group is English.

Prior knowledge of the following topics is a big plus:

• Relevant publications (and/or M.Sc. thesis) on the above-mentioned research topics
• Programming Microcontrollers and Interfacing Sensors
• Machine Learning Algorithms and Deep Neural Networks
• C/Assembly Language
• PCB Design and Basic Electronics, FPGAs

Expression of Interest

If you are interested in applying for the scholarship, please contact Assoc. Prof. Kasim Sinan Yildirim via email (kasimsinan.yildirim@unitn.it and [Ph.D.-TinyML-2023] as the subject of the email) by sending your CV (including three references and your previous publications).

Only the shortlisted candidates will be contacted for further steps.

• The tentative starting date of the Ph.D. scholarship is 01/11/2023. To get more information, visit the following links:
  • Information on becoming a Ph.D. student in UNITN.
  • Information on doctoral scholarships.

The Department of Information Engineering and Computer Science (DISI)

The Department of Information Engineering and Computer Science (DISI) is a leading and fast-growing research institution, characterized by a young and international faculty and by a large, international student population. Indicators for scientific production place the department among the top in Europe. The Department and the Ph.D. school closely collaborate with and operate in, a rapidly growing research and innovation environment characterized by top-class research centers and an increasing number of industrial research labs, including the Italian co-location center of the European Institute of Innovation and Technology (EIT).

Trento

Trento is a vibrant city with a beautifully preserved historic center, consistently ranked among the best cities for quality of life in Italy. It offers a variety of cultural and sports opportunities all year-round, as well as excellent food and wine.