New Computer Architectures for Probabilistic Computer Vision

The Physical Computation Laboratory is seeking a PhD student to work on a project developing new **Computer Architectures for Probabilistic Computer Vision Applications**.

The Physical Computation Laboratory (http://physcomp.eng.cam.ac.uk/) is a new and growing research group that investigates how to use an understanding of the physical world and the flexibility of sensing systems to improve the efficiency of computing systems that interact with nature. We conduct fundamental research augmented with hardware and software prototypes to get the results of our research out into the world. We have created several open source hardware and software platforms that are being used by several research groups and have active research collaborations with several departments across the University of Cambridge (from the Department of Applied Mathematics and Theoretical Physics, to the Department of Zoology) and across the world (from Max Planck to MIT).

The research of the successful applicant will investigate microarchitectures for probabilistic machine learning in Computer Vision applications. The research will investigate new hardware designs and evaluate the efficacy of the proposed designs and implementations using measurements on state-of-the-art prototype hardware platforms. This will be augmented with analysis of the effect of computer vision sensor characteristics on computation-/ data-flow graphs of important computer vision algorithm kernels.

The candidate will join a dynamic and multidisciplinary team of four PhD students, three postdocs, four masters students, and several industrial collaborators.

A successful candidate should have very strong background and excellent grades in their completed undergraduate/masters courses on:

- Digital logic
- Computer architecture
- Digital signal processing
- Probability and statistics

Candidates should also have a strong working knowledge of:

- C/C++ and Python
- Assembly language programming for one or more RISC architectures (e.g., RISC-V, PowerPC, or ARM)
- LaTeX
- Git (and a demonstrable record of working with repositories hosted on GitHub)
- Unix
- Mathematica or Matlab

Familiarity with popular machine learning tools (Caffe2, TensorFlow, TensorFlow Lite, MXNet) is a plus.




To find out more about the Physical Computation Laboratory, see http://physcomp.eng.cam.ac.uk/

To find out more about the Electrical Engineering Division at the University of Cambridge, see http://ee.eng.cam.ac.uk/

Applications should be submitted via the University of Cambridge Graduate Admissions web pages at https://www.graduate.study.cam.ac.uk/courses/directory/egegpdpeg/apply, with Dr. Phillip Stanley-Marbell specified as the potential supervisor.

To find out more about funding options, please see https://www.graduate.study.cam.ac.uk/finance/funding/graduate-funding-competition