Ann Arbor, MI, United States of America

Heewoo Kim

This inventor holds 1 USPTO granted patent. Top assignee: University of Michigan. Active years: 2025.


% Patents Active = 100.0

Average Co-Inventor Count = 5.0

ph-index = 1


Company Filing History:


Years Active: 2025

Loading Chart...
1 patent (USPTO):Explore Patents

Title: Heewoo Kim: Innovator in Hybrid Memory Systems

Introduction

Heewoo Kim is a prominent inventor based in Ann Arbor, MI (US). He is known for his groundbreaking work in memory systems, particularly in the realm of recommendation models. His innovative approach addresses the limitations of traditional memory systems, paving the way for enhanced performance in modern applications.

Latest Patents

Heewoo Kim holds a patent for a "DRAM-RRAM hybrid memory system for recommendation models." This invention tackles the performance constraints of modern recommendation models, which often suffer from memory bandwidth limitations due to embedding layer reductions. The hybrid memory system combines DRAM and RRAM with Processing In Memory (PIM) capabilities. It presents a comprehensive optimization strategy that includes access-pattern aware mapping, compute complexity reduction, and selective PIM reduction to mitigate computation latency. Evaluations indicate that this hybrid memory system achieves significant improvements in performance, energy efficiency, and energy-delay product (EDP), with enhancements of 2.6×, 1.7×, and 4.4×, respectively, compared to a CPU baseline.

Career Highlights

Heewoo Kim is affiliated with the University of Michigan, where he contributes to cutting-edge research in memory systems. His work has garnered attention for its innovative solutions to complex problems in the field of computer engineering.

Collaborations

Heewoo collaborates with notable colleagues, including Haojie Ye and Trevor N Mudge. Their combined expertise fosters a dynamic research environment that drives advancements in memory technology.

Conclusion

Heewoo Kim's contributions to hybrid memory systems represent a significant leap forward in the field of recommendation models. His innovative patent showcases the potential for improved performance and efficiency in modern computing applications.

Profile summary based on public USPTO records.
Please report any incorrect information to [email protected]
Loading…