Share

BrainChip Showcases Compelling Benchmarks and Recommends Better Metrics for AI Devices at the Edge

 

Laguna Hills, Calif. – January 15, 2023 – A new white paper by BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial producer of ultra-low power, fully digital, event-based, neuromorphic AI IP, evaluates the current state of edge AI benchmarks and the need to continually improve metrics that measure performance and efficiency of real-world, power-conscious edge AI deployments.

Current industry benchmarks measuring edge AI inference performance have started to capture the challenges of edge devices operation over the traditional TOPS rating. The paper, “Benchmarking AI Inference at the Edge: Measuring Performance and Efficiency for Real-World Deployments,” recommends the additional factors required to holistically gauge the performance and efficiency needed to enable compelling, optimized AI applications for complex, multi-modal edge environments.

BrainChip’s paper examines the limits of conventional AI performance benchmarks; discusses balancing performance and power at the edge; compares performance and energy efficiency with the tinyML benchmarks from MLCommons, which have made a good start towards identifying use cases; and shows how edge AI inference performance and efficiency is maximized with Akida™. It illustrates how various factors like model size, load times and system bandwidth can play a significant part in the overall result but aren’t currently accounted for. This is an area where the consortia should, and are, actively collaborating to improve. But there is room for more.

“While there’s been a good start, current methods of benchmarking for edge AI don’t accurately account for the factors that affect devices in industries such as automotive, smart homes and Industry 4.0,” said Anil Mankar, Chief Development Officer at BrainChip. “We believe that as a community, we should evolve benchmarks to continuously incorporate factors such as on-chip, in-memory computation and model sizes to complement the latency and power metrics that are measured today.”

To learn more about the importance of balancing these important criteria and better understand how BrainChip’s unique approach of event-based, neuromorphic design delivers compelling results, interested parties can download the white paper at www.brainchip.com

 

About BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY)

BrainChip is the worldwide leader in edge AI on-chip processing and learning. The company’s first-to-market, fully digital, event-based AI processor, AkidaTM, uses neuromorphic principles to mimic the human brain, analyzing only essential sensor inputs at the point of acquisition, processing data with unparalleled efficiency, precision, and economy of energy. Akida uniquely enables edge learning local to the chip, independent of the cloud, dramatically reducing latency while improving privacy and data security. Akida Neural processor IP, which can be integrated into SoCs on any process technology, has shown substantial benefits on today’s workloads and networks, and offers a platform for developers to create, tune and run their models using standard AI workflows like Tensorflow/Keras. In enabling effective edge compute to be universally deployable across real world applications such as connected cars, consumer electronics, and industrial IoT, BrainChip is proving that on-chip AI, close to the sensor, is the future, for its customers’ products, as well as the planet. Explore the benefits of Essential AI at www.brainchip.com.

 

Follow BrainChip on Twitter: https://www.twitter.com/BrainChip_inc

Follow BrainChip on LinkedIn: https://www.linkedin.com/company/7792006

 

###

 

Media Contact:

Mark Smith

JPR Communications

818-398-1424

 

Investor Contact:

Mark Komonoski

Integrous Communications

Direct: 877-255-8483

Mobile: 403-470-8384

mkomonoski@integcom.us

Related Posts

View all
  • Q&A with Peter Van Der Made, BrainChip Founder and Chief Technology Officer The semiconductor industry has long struggled to bypass Von Neumann bottlenecks, recalibrate Moore’s Law, and overcome the breakdown of Dennard Scaling. Advanced edge AI applications are fast approaching the limits of conventional silicon and cloud-centric learning models. With enormous amounts of targeted compute […]

    Continue reading
  • We are pleased to publish our second white paper, “Designing Smarter and Safer Cars with Essential AI.” This white paper presents a systems-architectural model that sequentially leverages multiple Akida-powered smart sensors and Akida AI SoC accelerators to efficiently capture and analyze inference data within designated regions or ROI. Click here to read how we’re addressing the automotive […]

    Continue reading
  • In collaboration with the GSA, we have developed a whitepaper on “Edge AI” to address the needs of latency, power, and security for AI applications at the edge. With smart sensors proliferating and the number of edge-enabled IoT devices expected to hit 7.8 billion by 2030, the semiconductor industry needs to more effectively address the […]

    Continue reading
  • By Brien M. Posey Previous generations of artificial intelligence and machine learning chips were useful, but their time is rapidly running out. The constraints on power and bandwidth imposed by edge devices mean that it’s time for a new paradigm, a new design that fulfills the promise of AI and ML at the edge. It’s […]

    Continue reading