Jump to content
  • Cognitive Computing

       (0 reviews)

    Tim Kannegieter

    Introduction

    Cognitive computing is an advanced form of data analytics that, broadly speaking, leverages the power of artificial intelligence and machine learning. With machine learning the computer learns from the data that it's monitoring. For example with unstructured data, a cognitive recognition system can be shown images that comply with, or do not comply with, the condition that we're looking for. The cognitive software then learns how to recognise the difference between them.

    The same is true for analysing more structured data. For example, when looking for patterns in sensor data that indicate potential failure, past data of failures is fed to the cognitive software which analyses it to discover correlations that is then used in rule based systems to detect and predict failure in the future.

    Cognitive systems stand in contrast to more traditional programmed systems, and proponents claim is it far better suited to IOT because the large volume and multivariate nature of the data. IoT data can be sourced not only from a huge range of traditional sensors (covering for example, vibration temperature etc) not just from the equipment itself but potentially from other sources such as weather data and even social media data as people reacting to conditions around them.

    This means the data has many potential ways of being analysed. IoT is not necessarily a case of knowing what question to ask, doing the analysis and coming up with an answers. Increasingly, there is so much data coming in that it is difficult to even know what questions to ask. Cognitive systems learn as they look at the data and they find insights the analyst may not even be suspecting.

    Examples

    One example of the application of cognitive computing in IoT is in health care for the elderly in their own homes (courtesy IBM). Asking the elderly to wear sensors is problematic because they may not raise an alert when they should or they alert when they shouldn’t and people stop wearing them after a while. An alternative approach is to instrument other things in the house such as fridge doors, light switches, bathrooms, movement sensors, and maybe infrared sensors etc. The cognitive software can thenbuild up an understanding of what normal looks like. When something abnormal happens, the system can then raise an alert and make a call to the emergency services.

    Another example (IBM working with Metronics in the US) is helping diabetic patients, manage their blood sugar levels. When blood sugar goes extremely low it triggers hypoglycemia and can cause a person to go unconscious or even potentially die. Cognitive computing capabilities are being used to detect patterns that indicate likely hypoglycemia two to three hours in advance, which allows the patient to take corrective action, eat some food, adjust the level of exercise and other things to avoid the hypoglycemia occurring at all.

    Drivers

    The growth of cognitive computing is being driven by many of the same factors as IoT in general, but in particular the growth in pervasive connectivity and cloud computing. The ability to process data in the cloud is bring many advanced analytical capabilities to bear on applications in the field, that previously was not possible.

    In terms of uptake, the growth of the IoT is allowing many organisations to consider IoT for the first time. As organisations install more sensors and instrument more things, they often getting more data than they have ever had before and are in a position to experiment and innovate.

    Challenges

    A challenge in cognitive computing is knowing what data to collect. Storage of data is not without cost and ideally there would be a systematic ways of determining what the sensible parameters in any given context may be useful for a cognitive system. However, it is dangerous to assume upfront what the data will reveal. A general rule may be to gather more data than less in the early phases when things are being learned and then pair back data collection when rules have been established. Fortunately, IOT data from traditional sensors is usually quite compact so usually the problem only arises when richer media is being used. Another approach is to use edge computing principles to push processing capability to devices like routers and switches and even hard-wired sensors, rather than have to have all the data sent across the cloud.

     

    Sources:



    User Feedback

    Recommended Comments

    There are no comments to display.



    Guest
    You are commenting as a guest. If you have an account, please sign in.
    Add a comment...

    ×   Pasted as rich text.   Restore formatting

      Only 75 emoji are allowed.

    ×   Your link has been automatically embedded.   Display as a link instead

    ×   Your previous content has been restored.   Clear editor

    ×   You cannot paste images directly. Upload or insert images from URL.


×