Perceive CEO, Steve Teig, will discuss the current industry focus on compressing weights to lower the memory requirements of neural networks – and why it’s necessary to investigate means to reduce activation memory as well. Although there are many schemes used to reduce weights, they often require compromises such as lower precision or smaller networks. Steve proposes that the compression of activations is the next step towards mitigating these issues. During his talk, he will introduce novel strategies in activation compression that could enable large networks to run efficiently on memory constrained hardware.

To learn more about this topic, please attend Steve’s talk at Embedded Vision Summit 2022.