Mehdi Ahmadi

Title: Mehdi Ahmadi: Innovator in Neural Network Pruning

Introduction

Mehdi Ahmadi is a notable inventor based in Montreal, Canada. He has made significant contributions to the field of neural networks, particularly in the area of filter pruning. His innovative approach aims to enhance the efficiency of convolutional neural networks.

Latest Patents

Mehdi Ahmadi holds a patent for "Methods, systems, and media for random semi-structured row-wise pruning in neural networks." This patent describes methods and systems for pruning rows of weights from kernels of filters in a convolutional layer of a convolutional neural network. The process involves using a pseudo-randomly-generated row pruning mask to optimize the network's performance. The convolutional neural network is trained to execute specific tasks using the pruned filters, which include only the rows of weights that remain after pruning. This method can be repeated multiple times to select the best-performing row pruning mask for deployment in processing systems, thereby improving computation time through the use of parallel hardware computation units.

Career Highlights

Mehdi is currently employed at Xfusion Digital Technologies Co., Ltd., where he continues to develop innovative solutions in the field of artificial intelligence. His work focuses on enhancing the capabilities of neural networks, making them more efficient and effective for various applications.

Collaborations

Some of Mehdi's coworkers include Vanessa Courville and Mahdi Zolnouri. Their collaborative efforts contribute to the advancement of technology at Xfusion Digital Technologies.

Conclusion

Mehdi Ahmadi is a pioneering inventor whose work in neural network pruning is shaping the future of artificial intelligence. His contributions are vital for improving the efficiency of convolutional neural networks, making significant strides in the field.