This video walks through the results of Numenta’s technology demonstration that shows 50x performance improvements on inference tasks in deep learning networks without any loss in accuracy.
The advances are achieved by applying a principle of the brain called sparsity. We compared sparse and dense networks by running our algorithms on Xilinx FPGAs (Field Programmable Gate Array) for a speech recognition task using the Google Speech Commands (GSC) dataset. Using the metric of number of words processed per second, our results show that sparse networks yield more than 50x acceleration over dense networks on a Xilinx Alveo board.
Additionally, we show the GSC network running on a smaller chip where dense networks are too large to run, enabling a new set of applications that rely on low-cost, low-power solutions.
Lastly, we show that the sparse networks use significantly less power than the most efficient dense network.
This proof-of-concept demonstration validates that sparsity can achieve significant acceleration and power efficiencies for a variety of deep learning platforms and network configurations, while maintaining competitive accuracy.
For more technical details of this technology validation, view the white paper here. Numenta is working with strategic partners to commercialize this technology. For information, contact firstname.lastname@example.org.