Location History:
- Ontario, CA (2012)
- Ottawa, CA (2018 - 2019)
Company Filing History:
Years Active: 2012-2019
Title: Innovations of Keith W. Campbell
Introduction
Keith W. Campbell is a notable inventor based in Ottawa, Canada. He has made significant contributions to the field of machine learning optimization, particularly in the context of graphical processing units (GPUs). With a total of 3 patents to his name, Campbell's work is recognized for its innovative approaches to enhancing computational efficiency.
Latest Patents
One of Campbell's latest patents is titled "Pipelined approach to fused kernels for optimization of machine learning workloads on graphical processing units." This method focuses on optimizing machine learning workloads on GPUs by identifying computations that exhibit generic patterns commonly found in machine learning processes. The approach includes hierarchical aggregation that spans the memory hierarchy of the GPU, allowing for the maintenance of partial output vector results in shared memory. Another significant patent under his name also revolves around the same theme, emphasizing the use of optimized fused GPU kernels to exploit temporal locality for data-flow dependencies. This method further enhances the efficiency of GPU kernel launch parameters, maximizing thread occupancy while minimizing atomic writes to GPU global memory.
Career Highlights
Keith W. Campbell is currently employed at International Business Machines Corporation (IBM), where he continues to push the boundaries of technology in machine learning and GPU optimization. His work has been instrumental in advancing the capabilities of GPUs in handling complex machine learning tasks.
Collaborations
Throughout his career, Campbell has collaborated with talented individuals such as Arash Ashari and Matthias Boehm. These collaborations have contributed to the innovative solutions he has developed in the field of machine learning.
Conclusion
In summary, Keith W. Campbell is a distinguished inventor whose work in optimizing machine learning workloads on GPUs has led to significant advancements in the field. His contributions are not only valuable to IBM but also to the broader technological landscape.