A recent blog by CISCO speaks of how QCM – Quantitative Complexity Management – is used in anomaly detection in highly complex systems and contexts. In particular, it speaks of QCM as being able to detect the existence of an anomaly without having seen it before. Such capability goes beyond Machine Learning in that QCM does not need to learn to recognize anything – it simply indicates sudden complexity jumps or gradients which are rarely a good omen.
Enhancing cybersecurity can be taken a step further. Imagine if routers, the key elements of our networks, were able to perform QCM-based anomaly detection. There are over 750 million routers scattered around the globe. Imagine if each one was able to do anomaly detection locally (this is called edge computing). Imagine what it would accomplish in terms of stopping, or debilitating cyber attacks before they have a chance to spread across the network. Imagine what it would do in terms of network resilience.
Clearly, this is a very futuristic vision and may take decades before it is implemented. However, innovation always starts with an idea and who knows where and by whom it will be implemented first.
Artificial Intuition 
Hi Jacek,
I cannot agree more …
futuristic vision and may take decades before it is implemented. However, innovation always starts with an idea !!!
And like many are getting so fed up with so many many many more hacks and wasted time to attacks and loss of personal data… wittingly or carelessly or unwittingly…
Do you think we can be enlightened with the next logical steps… how anomaly detection locally can be implemented via HW(QCM chip?) or SW App not necessarily on CISCO routers.. who might be having them in place soon… but also by the gamut of other routers out there namely by Big Blue… ASUS… nVidia…. Etc…..
Regards,
Nic
LikeLike