deep learning

February 27, 2018
Gordon Cooper

Optimizing power and performance trade-offs in CNN implementations for embedded vision

High-performance vision-processing algorithms need optimized CNN engines to deliver the right performance within the power budget of embedded applications.
May 23, 2016
Road signs used to train IDSIA neural network

Bit width tweaks point way to practical deep learning

In both data centers and automobiles deep learning is taking hold. But it is a technique that challenges conventional microprocessors, leading system designers to look at alternative architectures for acceleration.

PLATINUM SPONSORS

Synopsys Cadence Design Systems Mentor - A Siemens Business
View All Sponsors