Cupertino, CA, March 28, 2023 – Aizip, Inc., a leader in artificial intelligence (AI) for applications on the Internet of Things (IoT), also known as tiny machine learning (TinyML), announced today the Foundation Model for TinyML applications.
Over the past few years, foundation models have fundamentally reshaped AI system-development and many other domains, and we’re only at the beginning of this paradigm shift. Foundation models are trained with a vast quantity of data at scale, resulting in a model that can be adapted to a wide range of downstream tasks. GPT and Stable Diffusion, which have drawn worldwide attention since their release, are two examples of foundation models in the large language model (LLM) and AI-generated content (AIGC) space.
Aizip began exploring foundation models in 2020, developing its customized Foundation Model for TinyML (FMT). Since summer of 2021, Aizip has used FMT for automatic labeling, leading to tremendous savings in time and costs over human labeling. With the steady improvement of FMT, in 2022 Aizip also began to use it for robustness-enhancement in TinyML applications.
“We’re thrilled to see the significant impact of FMT on our TinyML model design and development process,” said Aizip Chief Scientist Brian Cheung. Cheung has led this project from the beginning. Data collection, curation, and labeling have challenged many IoT applications, and the manual process is costly and time-consuming. FMT brings TinyML data processing to a new stage and provides tremendous value. After two years of evolution, TinyML auto-labeling is now powerful enough to meet the requirements of these applications.
The robustness built into Aizip’s modes is a key factor of Aizip’s success in this space. FMT also plays a role, and has led to multiple production-quality models shipping in volume. “FMT has brought enormous value to our engineering development” noted John Rea, Chief Operations Officer overseeing all Aizip projects. “We continue to invest in this area and keep building on our technology leadership.”
Since its founding two-and-a-half years ago, Aizip has pioneered multiple areas of TinyML and established itself as a leader in this fast-growing market. Building on its momentum of volume shipping, Aizip is expanding in team and in product offerings. Aizip provides TinyML models to its IC partners and to module and system customers worldwide, and is committed to developing low-cost and high-performance intelligence for the benefit of all. For more information, please contact info@aizip.ai.
* * *
About Aizip, Inc.
Aizip, Inc., develops AI models for edge applications. Based in Silicon Valley, Aizip provides models with superior performance, quick turnaround time, and excellent ROI. These models can be used in a wide range of applications for an intelligent, automated, and connected world. For more information, visit www.aizip.ai.