Skip Navigation
Search
Back to Search Innovations
System, Method, and Accelerator to Process Convolutional Neural Network Layers
A CNN hardware accelerator that improves energy efficiency ideal for architectures that focus on the dataflow across convolutional layers
Source: Bartosz Kwitkowski, unsplash.com/photos/SJ5TmRRSM1U, Unsplash Licence.

Background

Deep convolutional neural networks (CNNs) are rapidly becoming the dominant approach to computer vision and a major component of many other pervasive machine learning tasks.

Technology

Developed is a novel CNN hardware accelerator with a new architecture and design methodology. Modified is the order in which the original input data are brought on to the chip. The design approach is a pyramid-shaped multi-layer sliding window, allowing effective on-chip caching during evaluation. Caching in turn reduces the off-chip memory bandwidth requirements.

Advantages

The proposed technology is an improvement in energy efficiency by minimizing data movements and improving performance.

Application

CNN accelerator architectures that focus on the dataflow across convolutional layers.

Inventors

Michael Ferdman, , Computer Science
Peter Milder, , Electrical Engineering
Monaj Alwani, student, Computer Science

Licensing Potential

Development partner - Commercial partner - Licensing

Licensing Status

Available for License. Stony Brook University seeks to develop and commercialize, by an exclusive or non-exclusive license agreement and/or sponsored research, with a company active in the area.

Licensing Contact

Donna Tumminello, Assistant Director, Intellectual Property Partners, donna.tumminello@stonybrook.edu, 6316324163

Patent Status

Patent application submitted - Provisional patent

US Provisional Filed

Tech Id

8826