Pat Lasserre

Title: Pat Lasserre: Innovator in In-Memory Computing

Introduction

Pat Lasserre is a prominent inventor based in San Jose, CA, known for his contributions to in-memory computing technologies. He holds a total of 3 patents that focus on enhancing the efficiency of neural networks through innovative memory solutions.

Latest Patents

One of Pat's latest patents is centered on in-memory matrix multiplication and its application in neural networks. This invention describes a device for in-memory vector-matrix multiplication that includes a memory array and in-memory logic. The memory array consists of at least two sections that store a multiplier matrix and receive an input multiplicand arranged in a vector. The design ensures that the operands of the vector-matrix multiplication are located in the same column of the memory array. Each section can be a volatile memory array, a non-volatile memory array, a destructive memory array, or a non-destructive memory array. The in-memory logic computes the output of the vector-matrix multiplication using the stored input vector and multiplier matrix. The memory array can be of various types, including RAM, DRAM, SRAM, Re-RAM, ZRAM, MRAM, and Memristor. Another significant patent involves a method for in-memory computation of a neural network, which includes storing the matrix of weights in an associated memory device, receiving an input vector, and computing the output of the network using the input and weights.

Career Highlights

Pat Lasserre is currently employed at GSI Technology, Inc., where he continues to develop cutting-edge technologies in the field of memory computing. His work has significantly impacted the efficiency and performance of neural networks, making strides in how data is processed and utilized.

Collaborations

Pat collaborates with Avidan Akerib, contributing to advancements in their shared field of expertise.

Conclusion

Pat Lasserre's innovative work in in-memory computing and neural networks showcases his commitment to advancing technology. His patents reflect a deep understanding of memory systems and their applications in modern computing.