Company Filing History:
Years Active: 2020-2024
Title: Innovations by Shabarivas Abhiram
Introduction
Shabarivas Abhiram is a notable inventor based in Mountain View, CA, with a significant contribution to the field of hardware architecture and convolutional neural networks. He holds a total of seven patents, showcasing his expertise and innovative spirit in technology.
Latest Patents
One of his latest patents is titled "Low power hardware architecture for handling accumulation overflows in a convolution operation." This invention focuses on a convolver unit that computes a running total by summing dot products during an accumulation cycle. The accumulator transmits an overflow indicator to a controller when the running total exceeds the maximum or minimum value, ensuring efficient operation. Another significant patent is "Low power hardware architecture for a convolutional neural network." This patent introduces a dynamic data quantization scheme that minimizes power consumption in systems implementing convolutional neural networks. It details how quantized representations of activation values and kernel parameters can be efficiently processed to enhance performance.
Career Highlights
Shabarivas Abhiram is currently employed at Recogni Inc., where he continues to develop innovative solutions in hardware architecture. His work has significantly impacted the efficiency and performance of convolutional neural networks, making strides in the field of artificial intelligence.
Collaborations
He has collaborated with notable colleagues, including Eugene M Feinberg and Gary S Goldman, contributing to various projects that push the boundaries of technology.
Conclusion
Shabarivas Abhiram's contributions to hardware architecture and convolutional neural networks exemplify his innovative spirit and dedication to advancing technology. His patents reflect a commitment to improving efficiency and performance in complex systems.