Page 96 - AI Standards for Global Impact: From Governance to Action

AI Standards for Global Impact: From Governance to Action

for power and compute efficiency, compact AI runtimes and frameworks, hardware virtualization,
and AIOps for AI-enabled applications.

AI models capable of running efficiently on edge devices need to be reduced considerably
in size and compute while maintaining similar reliable results. This process, often referred to
as model compression, involves advanced algorithms like neural architecture search (NAS),
transfer learning, pruning, and quantization. Model optimization should begin by selecting or
designing a model architecture specifically suited to the device’s hardware capabilities, then
refining it to run efficiently on specific edge devices. NAS techniques use search algorithms to
explore many possible AI models and find the one best suited for a particular task on the edge
device. Transfer learning techniques train a much smaller model (the student) using a larger
model (the teacher) that’s already trained. Pruning involves eliminating redundant parameters
that don’t significantly impact accuracy, and quantization converts the models to use lower-
precision arithmetic to save on computation and memory usage.

Some of the most popular AI models for vision applications—are designed to be extremely
efficient at these calculations. But in practice, these models do not always run well on the AI
chips inside our phones or smartwatches. This is because real-world performance depends
on more than just math speed—it also relies on how quickly data can move around inside the
device. If a model constantly needs to fetch data from memory, it can slow everything down,
no matter how fast the calculations are.

13�4 Edge AI use case implementations

13�4�1 Addressing untreated hearing loss

AI is rapidly transforming hearing technology. femtoAI addressed some of the biggest factors
contributing to the prevalence of untreated hearing loss with SPU-001 processors and deep
learning noise reduction algorithms by a.) solving the number 1 reported performance issue
with hearing aids, speech quality in noise, b.) reducing the stigma by enabling smaller, more
discrete form factors, and c.) reducing cost by bringing technologies to low cost OTC hearing
aids for an affordable price.

femtoAI developed an AI-driven chip technology for hearing aids, earbuds, and other audio
devices, using a method that mimics the brain’s efficiency to enable AI processing on low-power
devices. The AI chip optimizes speech separation and noise reduction while maintaining battery
efficiency – critical for hearing devices that operate within strict power constraints. Hearing
aids are becoming an arena for cutting-edge AI innovation. With AI-driven real-time speech
enhancement, modern hearing aids can now improve the clarity of conversations in noisy
environments, addressing one of the biggest barriers to adoption. As AI capabilities advance,
hearing aids are shifting from simple preset sound adjustments to adaptive, real-time sound
processing, providing a more personalized and effective hearing experience.

13�4�2 MountAIn Edge AI infrastructure in remote environments

In recent years, developments in IoT, AI, and satellite communications have been providing
data intelligence in ways that previously seemed inconceivable. But these technologies remain
somewhat siloed. While the potential opportunities in bringing these technologies together

84