Connected Devices have become a prominent trend this past decade. This is easily apparent in how every other new device produced today is engineered to be a connected device. There seems to be quite a bit of confusion around the concept though, as most of these products tend to be marketed as smart devices rather than as connected devices.

But though the terms ‘connected’ and ‘smart’ are frequently used interchangeably, there remains a significant and important difference between them. A connected device is only one tiny node in a massive Internet of Things. The ingredients that promise to bring these connected devices together into the revolutionary future we imagine, are artificial intelligence and machine learning.

Connected Devices vs Smart Devices

Connected devices are not a new phenomenon. Any device which can interact with our world and is connected to the internet is referred to as a connected device. They are merely devices that can transmit data and be operated remotely. A simple example would be a light bulb that you can switch on or off using your phone.

Smart devices, on the other hand, need to be context-aware. They should be able to react appropriately to changes in their circumstance or environment. A motion-activated light bulb can be considered a basic example of a smart device. This particular example also highlights how smart devices need not necessarily be connected, although it is worth noting that most smart devices today are also designed as connected devices.

From Smart Devices to Intelligent Systems

A smart device can sense its environment and react to triggers in a pre-programmed manner. This is achieved with the help of user-defined event-action rules and decision trees. These pre-programmed decision-making rules can only factor in a limited number of environmental changes.

In the real world, however, there can be any number of unpredictable events and changes. In order for a device or a system to be able to react to these changes in an appropriate manner, it needs to be able to learn from its environment. A system that can learn from patterns in its environment and make its own decisions based on those patterns would qualify to be called an Intelligent System.

Extending the earlier example further, in order for the motion-activated light bulb to exhibit intelligence, it should be able to stay switched off or get switched on depending on whether a person is simply tossing around on their bed, whether the person’s cat decided to take a midnight stroll, or whether the person themself is getting up for a glass of water.

The Difference that Machine Learning can make

With the IoT boom that started in the 2010s, there has been a huge surge in the number of connected devices available in the world today. These devices come with a variety of sensors and are constantly capturing and transmitting all kinds of data. This has resulted and will continue to result in the generation of humongous volumes of data. (One IDC report predicts that by 2025, connected devices will generate 79 Zettabytes of data — that is 79,000,000,000,000,000,000,000 bytes!)

The next generation of devices needs to be able to analyze these dense sets of raw data and extract relevant information from them. Machine learning algorithms facilitate this by being able to recognize patterns and identify anomalies in these complex and high-volume streams of data. These patterns and anomalies can then be used for decision making in the form of actionable triggers.

Using Machine Learning will allow devices and systems to become truly automated and operate with minimal or no manual intervention. This capability can be harnessed in so many different ways. Machine learning can detect the electricity and water consumption patterns and send alerts to help reduce wastage or ensure timely maintenance. It can be used to analyze consumption patterns and predict demand. It can be used to automate quality checks and track defects in an industrial setting.

Intelligent and Connected Systems of the Future

The true potential of connected devices can only be unlocked when they are integrated with artificial intelligence and machine learning to create an intuitive and efficient user experience. Over time, we can also expect to see complex systems of connected devices adaptively learning from each other.

Not too far in the future, we can imagine a scenario where your phone checks your morning meetings on your calendar, factors in the time it will take you to reach your workplace based on your route and the prevailing traffic conditions, and adaptively sets your alarm, switches on your lights, and starts your coffee machine — all to ensure you wake up at exactly the right time every morning!