3D circuit-technology co-exploration of convolutional neural networks

Leuven - PhD
|
Meer dan twee weken geleden

Convolutional and Deep Neural Networks have received much attention and investment from the research community as well as industry in recent years but a strong push exists towards reducing the cost and especially energy-efficiency which requires pushing the boundaries of classical architecture design paradigms and to co-optimize them together with circuit design and technology implementation

Solliciteer

Convolutional and Deep Neural Networks have received much attention and investment from the research community as well as industry in recent years owing to the highly accurate performance in certain classes of machine perception tasks. This coupled with the ever-increasing demand for smart systems is driving the need for continuous improvement in performance while requiring the technology to be cheaper and energy-efficient and thus
more widely available in portable/nomadic applications. A large part of these advancements have relied on the improvement of computing hardware over the last few decades and the majority of CNN/DNNs today run on high-performance computing platforms, like multi-core CPUs and GPUs. However a push exists towards reducing the cost and especially energy-efficiency. This has given rise to a new and interesting research problems that require pushing the boundaries of classical architecture design paradigms and to co-optimize them together with circuit design and technology implementation.

In this PhD proposal we want to go into this architecture-circuit-technology co-optimisation direction.  A pipelined data-flow scheme which eliminates the need for costly local (SRAM) memory accesses during the tensor convolution execution will be used as basis. But many ways exist to project that data-flow into a cost/energy-effective architecture and circuit. We want to explore that broad search space for the context of (3+1)-D convolutions. We also want to exploit emerging technology options which support effective use of the 3rd scaling dimension.  That will enable potentially strong gains in the interconnections which typically dominate the realisation of large processing networks like CNN/DNNs.

 

Required background: Electrical Engineering

Type of work: 30% architecture design, 40% circuit design, 30% implementation and simulation

Supervisor: Marian Verhelst

Daily advisor: Marian Verhelst

The reference code for this position is 2020-053. Mention this reference code on your application form.
Chinese nationals who wish to apply for the CSC scholarship, should use the following code when applying for this topic: CSC2020-20.

Solliciteer

Share this on

truetrue

Deze website maakt gebruik van cookies met als enige doel het analyseren van surfgedrag, zonder enige commerciële insteek. Lees er hier meer over. Lees ook ons privacy statement. Sommige inhoud (video's, iframes, formulieren,...) op deze website zal pas zichtbaar zijn na het accepteren van de cookies.

Accepteer cookies