This inventor holds 3 USPTO granted patents. Top assignee: Magic Leap, Inc.. Active years: 2019-2023.
Company Filing History:
Years Active: 2019-2023
Title: Ashkan Aliabadi: Innovator in Convolutional Neural Networks
Introduction
Ashkan Aliabadi is a prominent inventor based in Santa Clara, CA (US). He has made significant contributions to the field of convolutional neural networks, holding a total of 3 patents. His work focuses on enhancing the efficiency of data layouts in neural network implementations.
Latest Patents
One of Ashkan's latest patents is titled "Efficient data layouts for convolutional neural networks." This patent discloses systems and methods for the efficient implementation of a convolutional layer within a convolutional neural network. In one aspect, the weight values of kernels in a kernel stack can be reordered into a tile layout with tiles of runnels. Additionally, pixel values of input activation maps can be reordered into an interleaved layout comprising a plurality of clusters of input activation map pixels. The output activation maps can then be determined using these clusters of input activation map pixels and kernels tile by tile.
Career Highlights
Ashkan Aliabadi is currently employed at Magic Leap, Inc., where he continues to innovate in the field of technology. His work has been instrumental in advancing the capabilities of convolutional neural networks, making them more efficient and effective.
Collaborations
Some of Ashkan's notable coworkers include Gregory David Roberts and Gregory C Roberts. Their collaboration has contributed to the development of innovative solutions in their field.
Conclusion
Ashkan Aliabadi is a key figure in the advancement of convolutional neural networks, with a focus on efficient data layouts. His contributions through patents and his work at Magic Leap, Inc. highlight his commitment to innovation in technology.