Reducing Compute Resources in Neural Networks - CitC Episode 266

Intel CitC

Helen Kim from MaxLinear (previously NanoSemi, Inc.) joins host Jake Smith to talk about reducing compute resources to achieve target accuracies in deep neural networks. Helen goes into detail about MaxLinear’s Augmented Neuron technology, which mathematically augments neural networks to reduce memory usage and latency. Jake and Helen discuss how Intel’s oneDNN and other tools are making AI advancements easier for partners and how the future of 5G will impact the larger industry. For more information, visit: https://www.nanosemitech.com/benchmarks-show-maxlinears-augmented-neuron-reduces-resnet50-cost-by-2x/ Follow Jake on Twitter at: https://twitter.com/jakesmithintel