Electronics industry: can Europe take a place thanks to embedded AI and its promising applications in audio? – Silicon

Europe has a real opportunity to get back in the game by accelerating RISC-V technology and positioning itself as a leader in embedded AI.

The design of objects equipped with embedded artificial intelligence was until now inaccessible due to technological limitations. Thanks to the arrival of RISC-V, new generation open source processors, this is now possible. Carried by expected applications in audio, embedded AI could constitute a real opportunity for Europe in the electronics industry market, currently in the midst of a struggle for Sino-American hegemony.

Next-gen processors are shaking up the electronics industry

Integrated circuits are at the heart of everyday electronic objects, whether our smartphones or headphones. The “ISA” market, the architectures used to program processors, has historically been dominated by ARM. But for the past five years – a short time in terms of ISAs – a new architecture, RISC-V, has generated strong enthusiasm among manufacturers.

To date, RISC-V processors are already integrated into 25% of the digital circuits being designed, even though they are still excluded from high-volume markets such as computers or smartphones. These markets could nevertheless also open up in the future, if we stick to Apple’s desire to “complete the software and hardware team that implements innovative RISC-V solutions” expressed in a recent offer employment.

The rapid progress of RISC-V is largely due to its open source availability, lower power consumption, and reliable security features. Because of these attributes, the possibilities for innovation in embedded artificial intelligence (AI) are seen as limitless on these new processors.

According to a study conducted by Counterpoint Research, the processor architecture market which was worth $5.2 billion in 2020 is expected to reach $8.6 billion by 2025, driven in particular by the growing demand for advanced processor applications. devices using artificial intelligence.

Embedded AI becomes a reality thanks to the innovation capacity of open-source

Through the collection of a large number of representative examples called “training data”, AI has made it possible to design adaptive functionalities. Until recently, advances in AI such as voice recognition mainly concerned software, which was almost always deported to the Cloud.

The arrival of applications requiring strict compliance with data, such as the IoMT (Internet of medical things), or a response in near real time, such as the braking systems of autonomous cars or hearing aids, changes the situation. In this context, it is impossible to turn to the Cloud, whose risks related to data transmission and processing times are too high.

AI must be embedded directly in components offering adequate memory capacity and computing power, beyond the reach of generic processors that are not efficient enough to perform real-time AI inference calculations. The challenge thus lies in the architecture, which places RISC-V at the center of the game due to the capacity for innovation offered by open source.

In this context, there is therefore no longer a priori any limit to the introduction of embedded AI in markets with high application potential such as, for example, that of audio.

AI opens up a vast field of possibilities for new audio applications

The audio market offers strong application potential for embedded AI. Based on statistical mathematics, the current on-board algorithms allow a variation of predefined settings to adapt performance to the environment and use. Thus, the noise cancellation (ANC) can adapt to the surrounding noise (that of the subway, the office or the street), or the level of gain of the hearing aids can adapt to the level of perceived voice.

The new generation of processors together with innovation in AI offer a new paradigm in the audio experience. The AI ​​algorithms come, by a game of learning from the user’s feelings, adapt the performance over time, to the point of personalizing the products to the person without any pre-adjustment being required. The voice signals and the musical flow can thus be adapted to each person’s habits and hearing, making it possible to change the use of the products.

Helmets are currently prohibited on bicycles but will, without a doubt, be recommended when they allow you to detect an audible alert in the hubbub of traffic. Long-awaited features will emerge, such as live audio translation, or hearing aids capable of adapting in real time to the user’s auditory profile and their environment.

A technological turning point is therefore underway, driven by hardware and software progress. It is important not to miss it. In a context of Sino-American hegemony in the manufacture of processors, with centralized production in Taiwan and GAFAM control of remote AI processing in the Cloud, Europe has a real opportunity to get back into the race. by accelerating RISC-V technology and positioning itself as a leader in embedded AI.

We would love to say thanks to the writer of this short article for this amazing content

Electronics industry: can Europe take a place thanks to embedded AI and its promising applications in audio? – Silicon


Find here our social media profiles as well as other pages that are related to them.https://www.ai-magazine.com/related-pages/