PhD in Digital Design for Efficient Embedded Machine Learning Processors

Catholic University of Leuven Kulak

Belgium

 

MICAS is a research group within the department of Electrical Engineering (ESAT) of the University of Leuven (KU Leuven). MICAS conducts research on the design of integrated circuits in nano-electronic technologies. This research is performed in close collaboration with the semiconductor industry and about 75% of the groups’ budget is through industrial collaborations. The staff members of MICAS are: Prof. Wim Dehaene, Prof. Georges Gielen, Prof. Bob Puers, Prof. Michael Kraft, Prof. Patrick Reynaert, Prof. Michiel Steyaert, Prof. Filip Tavernier and Prof. Marian Verhelst. In total, there are 79 Ph.D. researchers working within MICAS. As such, KU Leuven ESAT-MICAS is the largest academic circuit-level research group within Europe. MICAS tapes out dozens of chips every year, and possesses a top notch chip bonding, packaging and measurement lab. More information about the group and its activities can be found at http://www.esat.kuleuven.be/micas/

 
Project

We want to extend our team with a PhD on resource-efficient digital implementations of machine learning processors, focused around deep learning.

Recently deep neural networks, such as convolutional neural networks (CNNs) or long short-term memory (LSTM) networks have gained enormous popularity in the signal processing community. In the micro-electronics research domain this has sprouted attention on customized processors for efficient embedded deep neural network inference. Our team has published several of these state-of-the-art processors over the past few years.

We will build upon our work in the field of deep learning processors and hardware/algorithm co-optimization, which has demonstrated to save orders of magnitude on energy efficiency. Yet, we want to extend this work towards a.) new application domains (such as acoustics, depth sensing and sensor fusion); b.) online learning for user customization; c.) run-time adaptivity for increased efficiency.

In this project, we closely collaborate with researchers from KU Leuven’s algorithmic machine learning group, various international research groups and several partner companies.

 
Profile
  • Candidates must hold a Masters degree in Electrical Engineering (or equivalent), with a background in digital design.
  • Additional research / expertise in computer architectures, machine learning or chip tape out is a strong plus.
  • We are looking for a team player with the capability to work in an international research team.
  • Excellent proficiency in the English language is also required, as well as good communication skills, both oral and written.
 
Offer
  • An exciting research environment, working on the intersection between emerging research domains (machine learning; processor design; ultra-low power chip and system design).
  •  Interactions with many international befriended research teams, and companies, Collaborations with Imec.
  • A Ph.D. title from a highly-ranked university (after approximately 4 to 5 years of successful research).
  • A thorough scientific education, the possibility to become a world-class researcher.
  • A KU Leuven affiliation, one of the largest research universities of Europe.
  • The possibility to participate in international conferences and collaborations.
 
Interested?
For more information please contact Prof. dr. ir. Marian Verhelst, tel.: +32 16 32 86 17, mail: marian.verhelst@kuleuven.be.
You can apply for this job no later than August 06, 2019 via the online application tool
KU Leuven seeks to foster an environment where all talents can flourish, regardless of gender, age, cultural background, nationality or impairments. If you have any questions relating to accessibility or support, please contact us at diversiteit.HR@kuleuven.be.


In your application, please refer to Polytechnicpositions.com

FACEBOOK
TWITTER
LINKEDIN
GOOGLE
https://polytechnicpositions.com/phd-in-digital-design-for-efficient-embedded-machine-learning-processors,i3257.html">

baner1

baner10

baner11

baner12

baner14

baner2

baner3

baner4

baner5

baner6

baner8

baner9