Making the artificial intelligence of things a reality
Artificial Intelligence (AI) has been in the public consciousness for much longer. For many, the most memorable incarnation is HAL 9000, a fictional artificial intelligence character that first appeared in 1968 in Arthur C. Clarke’s Space Odyssey series. Happily, HAL 9000 bears very little resemblance to the reality of today’s AI systems, which have more straightforward intentions and almost always ‘live’ in massive, remote data centres. They are capable of absorbing, structuring and processing huge quantities of data, generating new insights based on sophisticated analytics algorithms and, in some still limited circumstances, streamlining decision making.
The Artificial Intelligence of Things (AIoT) is the point at which these two parallel trends converge; bringing with it the potential for a complex, intelligent network of decision making and analytical processes, collating and interpreting swathes of data that enable previously impossible use cases. Crucially, the AIoT will process this data on the device itself. Part of the genius of these devices is that the magic is happening in the palm of your hand, or in your immediate vicinity, independent of data centres and external networks.
Processing at the edge
You might ask why you need the AIoT, given the convenience and popularisation of the cloud. For many businesses, digital transformation has simply meant a shift of their services and networks into the cloud infrastructure, and it’s easy to assume that you could send data to and from that invisible resource without any issues.
The truth is that the cloud simply cannot keep up with the growing scale of the IoT. Communications infrastructure are bandwidth limited, data centres cannot deliver the compute. This results in a myriad of problems with network bandwidth, latency, compute scalability, energy management and security. Relying on existing cloud architectures is simply not a realistic option on anything more than a localised, private scale, which would in turn limit the scope and the potential for AIoT applications.
With Business Insider Intelligence suggesting that there will be 64 billion IoT devices as early as 2025 – an explosive increase of more than 600% on 2018’s 10 billion – the AIoT is in need of devices with incredible processing power at the very edge in order to meet users’ expectations.
Rather than sending data to and from colossal cloud networks, the AIoT sees processing power being moved towards or directly on to the devices generating the data itself. This takes the strain away from overloaded networks and power hungry and costly data centres groaning under the weight of requests, distributing the workload in a way that greatly improves performance.
However, there are clear financial difficulties here. For many businesses looking to take advantage of AIoT, the required investment in high-end CPUs to comprehensively cover their network endpoints is completely unrealistic. But without the ability to seriously boost processing power, AIoT is essentially a pipe dream.
The hardware behind it
As such, there’s a rapidly growing demand for chips that can enable high performance processing without a debilitating price tag. Any manufacturer looking to produce a processor that meets the needs of these businesses has a precarious balancing act to perform:
Cost: The processors that meet this need to be cheaper than the premium alternatives that many would assume were the only solution to the need for compute power. There’s a wealth of trip hazards to negotiate here, from the expenses of third-party hardware/software to the category of the processor and the components it requires to function.
Performance: Cheap cannot mean compromising on performance, or the entire operation becomes pointless. The minds behind the product need to be able to compensate for cost-cutting measures with sophisticated algorithms and architectures to ensure that the quality bar is met.
Versatility: Perfecting the split between price and performance is all well and good, but if the result is a chip purpose-built for one use – say, driving voice recognition – then the majority of the AIoT industry hasn’t been served a solution at all. You see, the big secret is that AIoT is not really a market at all, it is the sum of 100 markets, and 10,000 market segments – each with different needs. The endpoint solution must flexibly and affordably deliver the combination of compute classes (AI, DSP, control and IO) required by each of these segments.
From the micro to the macro, hardware that meets these three parameters will have a potentially revolutionary impact on the spread and scope of the AIoT.
And the impact can be enormous. For example, AIoT systems open up the possibility to improve quality of life through connected healthcare – IoT devices that are capable of monitoring things like heart rate or breathing patterns, could pre-emptively flag an incident before it occurs. Over time, the data from these incidents could be shared with a GP or hospital directly – a living record of your, or your family’s, health that helps to guarantee the most appropriate treatment from healthcare professionals.
Expand this a little further, and the smart home opens up opportunities for safety, convenience and automation that resemble something out of Futurama. Voice-controlled thermostats, lights, entertainment devices and more could offer total control without the need to even divert your attention, let alone open a menu.
Widen this further still, and a plethora of smart homes give birth to the smart city, offering the chance to reach new heights in convenience and efficiency. Imagine the saving in time and energy that would be delivered by being directed straight to an empty parking space in a busy city. Streetlights on quiet roads can be turned off until proximity sensors X metres away detect an oncoming car. Microphones can guide emergency services towards noises like gunshots or breaking glass.
No matter how far you want to take the concept of AIoT, its success depends on being able to drive one of the most impressive feats of electronics engineering the digital era has ever seen. If we can secure versatile, processors that pack the necessary punch at the right price, there is no reason why we can’t all be the beneficiaries of an intelligence revolution.
About the author:
Mark Lippett is the CEO of XMOS – www.xmos.com