Skip to content

Machine Learning

When two new and exciting technological innovations are paired together, the results can be ground breaking: the marriage between phones and copy machines brought us fax machines and satellites’ union with portable screens gave us GPS.

While today they may be ordinary everyday devices, at the time of their invention they were all cutting edge technology and a giant step in development of next generation electronics. Once again, we may just be at the precipice of something similar with IoT and machine learning.

In a similar way to the growth of smart phones in the late 2000s, IoT has come to represent a similar trend. While not excessively common amongst regular consumers right now, it is predicted to grow with intense speed over the next few years. Machine learning, on the other hand, has been a fascinating concept for decades, stemming from the idea of Artificial Intelligence (AI). While at least for now AI might not be a reality, it does not mean machine learning isn’t. Machines can be seen analysing data and learning from the bits of information users put in – visible in e.g. health apps, predictive statistics, and search engines – and by now, it’s nothing out of the ordinary.

Working together, the two are already bringing interesting products to market; smart light bulbs, for example, learn the users’ routine and know when to turn on, or smart kettles that can predict when you might want your coffee. It goes way beyond that though; self brewing hot drinks may sound like a novelty, but it is only a fraction of the potential that IoT and machine learning have when merged together. The Nest thermostat, for example, adjusts to the occupants’ waking and sleeping times and based on this regulates air temperature and heating. A more complicated variation is Jaguar’s intelligent car. It is able to recognise people by their smartphones, adjust seating and mirrors according to preference (which it has learnt), and walk the driver through their diary. The system is said to “study, predict, check and remind the car’s occupants.” Essentially, IoT and machine learning are an alternative to cookies, only on a larger scale.

Through long term development the combination can also further robotics or even aid in the study of AI. Some sources, such as technology news site Wired, claim IoT will not be able to develop any further without machine learning. IoT’s analysis solutions fail to reach the volumes set by IoT devices and to ensure data is analysed in real time, machine learning has to step in and take over.

Dependent on machine learning or not, it can be easily agreed that together they are set to be revolutionary. The expectations are high and judging by the results so far, we won’t be disappointed either.