Artificial intelligence gets a realtime local edge

When computing moved to the cloud, it brought a lot of advantages of managing data at scale. Computation of large sets of information sourced from several locations allowed experts to interpret data at scale with flexibility. Artificial intelligence (AI) played a critical role in helping understand patterns which could lead to business insights. 

In the new scenario, using AI in the cloud is not enough anymore. The rise of billions of connected devices mean that data has to be processed locally. Nearly every connected device now has processing power and needs information stored locally to be computed with context. Most importantly, this needs to happen in realtime with minimal delay.

A mobile phone or smart TV, for instance, must be able to do its own processing in realtime. For this the data and computing ability has to be done on the device. If the data is on the cloud and the device has to seek information constantly, there will be a delayed response. The delay or latency can be milliseconds, but even this much is too much in some cases. If the device is enabling a remote surgery or managing a car being driven, then a delay of milliseconds can negatively impact the outcome. 

This is where edge computing is combining with AI to ensure that connected devices can interpret data and deliver results much faster. Edge computing is the processing of information on a local device or platform closest to the application, effectively at the edge of a network. Cloud computing is remote and is deep inside the network. Now AI is being deployed to bring the advantage of high-level computing at local devices. 

 

Edge brings the advantage of generic data stored in the cloud to local devices for specific computing. While AI can manage patterns at scale on the cloud, it is also deployed on a device or sensor for localised computing and delivery of objectives. 

“In 2022, the market for the Internet of Things is expected to grow 18% to 14.4 billion active connections. It is expected that by 2025, as supply constraints ease and growth further accelerates, there will be approximately 27 billion connected IoT devices,” according to a report by IoT Analytics.

Edge AI computing will allow billions of such devices to process information and manage outcomes at far greater speed and efficiency. The local computing prowess boosted by AI will reduce latency and help users with constantly improving outcomes. 

As 5G networks begin to roll out across the world, the network connectivity will further accelerate the use of edge AI computing. 

Several start-ups and legacy companies in India are moving to deliver edge AI. An “Edge analytics application use-case spectrum is very broad, including wearables, smart homes, smart cities, autonomous cars and industry automation. The health industry stands to benefit greatly by using EA with wearables to monitor, identify issues and prescribe remedies affecting human health as early as possible,” according to an analysis by Wipro. “Detecting sounds such as that of a blast, a baby crying, or glass breaking will trigger an alert. Recognising specific sounds from within overlapping sound sources is a critical ask that AI can fulfil. Hands-free read and write facilities can be enabled by means of AI at the edge.”

Many other sectors will benefit from Edge AI. These include civic management, power grid efficiency, telecom networks, highways and traffic control. 

For an economy like India, where processing of data has to be at scale and also local, Edge AI applications would turn out to be very beneficial.