Real-World Examples of Hyperdimensional Computing for Industry / IIoT

Oct 6

4 min read

Introduction

Hyperdimensional computing (HDC) is emerging as a powerful tool in the realm of artificial intelligence and machine learning. This innovative approach to computing is now finding its way into real-world industrial applications, offering promising solutions to complex problems. Let's explore some concrete examples of how companies and institutions are leveraging HDC in practical settings.

To learn about HDC in action, please see:

Northrop Grumman Explores HDC in Aerospace

Aerospace giant Northrop Grumman is at the forefront of HDC exploration in the industry. The company has partnered with researchers to test HDC capabilities on their stochastic computing chip. This collaboration aims to showcase the power of HDC in hardware implementations, potentially revolutionizing aerospace technologies. By investing in HDC research, Northrop Grumman is paving the way for future advancements in aerospace computing and data processing.

Here is an excerpt of an article discussing Northrop Grumman's cooperation with UC San Diego:

"Overall, the team will have to develop new coding and decoding strategies, fast HD algorithms and efficient hardware. On this last point, the HyDREA team is collaborating with Northrop Grumman to test HD computing on the aerospace company’s stochastic computing chip." Source: UC San Diego’s Rosing Receives DARPA Grant to Explore Hyperdimensional Computing

Zscale Labs™ Already Uses Hyperdimensional Computing (HDC)

At the forefront of hyperdimensional technology applications is Zscale Labs™ (www.ZscaleLabs.com), a company leveraging hyperdimensional computing (HDC) in combination with neuro-symbolic AI (NSAI) to create advanced artificial intelligence systems. Zscale Labs™ utilizes the unique properties of high-dimensional spaces to develop AI that can learn and adapt in real-time, mirroring human cognitive flexibility.

Their approach enables the creation of AI systems with unprecedented efficiency, adaptability, and intuitive problem-solving capabilities. These systems are particularly well-suited for edge computing and resource-constrained environments due to their low power consumption and high-speed processing abilities. The robust performance of Zscale Labs™' HDC-based systems, even in noisy or uncertain conditions, makes them ideal for a wide range of real-world applications.

South Korean Team Commercializes HDC for Robotics

A team of researchers from South Korea, including members from Hanyang University and Daegu Gyeongbuk Institute of Science and Technology, has successfully commercialized an AI learning method using HDC. Their groundbreaking work has been applied to indoor autonomous driving robots, yielding impressive results. The HDC-powered robots achieved speeds 15 times faster than those using traditional deep learning methods. Moreover, this implementation came at a fraction of the cost and power consumption, operating at one-thirtieth the cost and one-twentieth the power usage of conventional approaches.

"HDC is considered a next-generation learning method that could replace the deep learning approach, which is often criticized for its high costs and significant power consumption in AI training and inference. "HDC mimics the brain’s computing method by not storing information in specific neurons but rather distributing it across multiple neurons. This system is designed using thousands of unique hyperdimensional vectors, which represent and combine all objects, concepts, and functions, allowing for rapid inference results through simple computations." - Business Korea reporting on Hanyang University's HDC

University of California San Diego Tackles COVID-19

The University of California San Diego (UCSD) has implemented HDC in an innovative application for COVID-19 wastewater surveillance. This use of HDC has significantly accelerated genomic sequence tracking, resulting in a 200-fold increase in DNA sequencing speed. The impact of this advancement was substantial, enabling early detection of 85% of COVID-19 cases on the UCSD campus. This real-world application demonstrates the potential of HDC in public health and rapid response scenarios.

The Future of HDC in Industry

While these examples showcase the potential of HDC in various sectors, it's important to note that many industrial applications are still in the early stages of development or adoption. As the technology matures and more companies recognize its benefits, we can expect to see a wider range of HDC implementations across different industries.

The unique advantages of HDC, such as improved speed, reduced power consumption, and lower costs, make it an attractive option for companies looking to enhance their AI and machine learning capabilities. As research continues and more real-world applications emerge, HDC could play a crucial role in shaping the future of industrial computing and automation.

In conclusion, hyperdimensional computing is moving beyond theoretical concepts and into practical, real-world applications. From public health surveillance to aerospace technology and robotics, HDC is proving its worth in diverse industrial settings. As more companies and institutions explore its potential, we can anticipate exciting developments and innovative solutions to complex industrial challenges in the years to come.

***

References:

• https://www.ucsd.edu/

• https://www.northropgrumman.com/

• https://www.hanyang.ac.kr/

• https://www.dgist.ac.kr/

• https://www.nature.com/articles/s41598-021-02077-6

• https://ieeexplore.ieee.org/document/9743766

• https://www.sciencedirect.com/science/article/pii/S2666307422000043

• https://www.mdpi.com/2079-9292/10/16/1895

• https://www.frontiersin.org/articles/10.3389/fnins.2021.675597/full