Hoai Linh Tran

Title: Hoai Linh Tran and Convolution Gradient Kernel Innovation.

Introduction

Hoai Linh Tran is an inventor associated with Markham, Canada, and Huawei Technologies Co., Ltd. Her patent work addresses methods for accelerating convolution operations, a key computation in machine learning and other data-intensive applications.

Latest Patents

Tran is credited with 1 patent: “Method and Apparatus for Generating Architecture Specific Convolution Gradient Kernels.” The invention describes receiving two data sets through an I/O interface, converting each into a format suited to a convolution functional unit, and loading the converted data into that unit. The unit processes the inputs and outputs a result in a third format, which a task scheduler receives. This approach is designed to support efficient, architecture-specific convolution processing.

Career Highlights

Tran’s listed patent reflects work on adapting data formats and coordinating computation within a convolution system. The patent identifies Huawei Technologies Co., Ltd. as her company affiliation.

Collaborations

The provided information lists Zichun Ye and Giancarlo Colmenares as Tran’s coworkers. Their names are included here as professional associations, without implying a specific role in the patent.

Conclusion

Hoai Linh Tran’s patent focuses on preparing and processing data efficiently for convolution operations, highlighting an innovation in specialized computing architecture.