Imec is the world-leading research and innovation hub in nanoelectronics and digital technologies. The combination of our widely acclaimed leadership in microchip technology and profound software and ICT expertise is what makes us unique. By leveraging our world-class infrastructure and local and global ecosystem of partners across a multitude of industries, we create groundbreaking innovation in application domains such as healthcare, smart cities and mobility, logistics and manufacturing, and energy.
As a trusted partner for companies, start-ups and universities we bring together close to 3,500 brilliant minds from over 70 nationalities. Imec is headquartered in Leuven, Belgium and also has distributed R&D groups at a number of Flemish universities, in the Netherlands, Taiwan, USA, China, and offices in India and Japan. All of these particular traits make imec to be a top-class employer. To strengthen this position as a leading player in our field, we are looking for those passionate talents that make the difference!
The Internet and Data Lab (IDLAB, http://idlab.ugent.be) research group is part of Ghent University, Belgium and a core research group of imec (http://imec.be) , the world-leading research and innovation hub in nano-electronics and digital technologies. IDLab performs fundamental and applied research in research areas such as machine learning and distributed intelligence for IoT.
Recently, imec has unveiled the world’s first self-learning neuromorphic chip based on OxRAM technology. Imec researchers are combining state-of-the-art hardware and software to design chips that features huge computing power while only consuming a few tens of Watts.
In this context, imec is building a team combining the analog and digital hardware design expertise in Leuven with machine learning (deep learning and neuromorphic paradigms) in Ghent. To further strengthen this team, we are looking for 2 PhD students.
Neuromorphic-accelerated deep learning algorithms and applications
Compared to today’s GPUs, neuromorphic hardware acceleration provides two major advantages, each with a few orders of magnitude: power efficiency and data throughput. Hence, this opens unique opportunities for novel types of applications that are deemed unrealistic today, in particular those applications that apply on-chip self-learning. The goal of this PhD is to develop suitable deep learning architectures and procedures for input data shaping for neuromorphic hardware. We will specifically focus on neural networks with external memory access and self-reconfiguring neural networks for on-chip learning.
Mapping deep learning algorithms to neuromorphic hardware architecture
Dominant models for neural network architectures and training procedures are highly optimized for training and inference on GPU-based clouds. Neuromorphic architectures leave the traditional Von Neumann architecture and combine processing with memory in a single place. The goal of this PhD is to optimize the mapping of neural network architectures and training procedures in terms of hardware-related metrics such as performance, power and area.
We are looking for candidates with a master degree in computer science, analog/digital electronics , embedded systems or microarchitectures. Knowledge of non-volatile-memory (NVM) technology fundamentals, deep learning or mapping algorithms (signal processing or machine learning) on processor architectures (GPU, DSP, FPGA) will be considered a strong plus. Candidates who are close to their master degree are also welcome to apply. Further, candidates are expected to have excellent communication skills, and to be able to work in a multidisciplinary team under a collaborative spirit.
We offer a full-time contract with various additional benefits.
Candidates should apply with their CV, including a resume of the master thesis and (if applicable) publication list). Please clearly indicate for which topic you area applying.