Tuesday, March 1, 2016

Cognitive Computing For All?

There seems to be great promise for cognitive computing, but what is it and how can it help individuals and business in ways that contribute to a thriving world? I'm going to cheat a bit and use a definition I found on the web that is a good enough starting point:

Cognitive computing is the simulation of human thought processes in a computerized model. Cognitive computing involves self-learning systems that use data mining, pattern recognition and natural language processing to mimic the way the human brain works.

I would like to add value by showing where cognitive computing adds to or extends the human experience. I believe that it can augment human thinking and actions safely, but once cognitive becomes autonomous (acting on it's own), then special care should be taken to put in safety policies, rules and constraints (boundaries). 

Assuming that all forms of big data(data, events, voice and video) can be consumed and organized for cognitive agents (Cogs), cognitive can bring about great benefits for organizations going digital. I've identified five major areas of cognitive contribution that seem to be contributing today. I will gather case studies over time to publish to show traction. The five areas of contribution are as follows:

Help Decide:

Cognitive is great for highlighting great alternatives for humans by consuming large amounts of sensory data and showing the way forward. This is great for augmenting knowledge workers in areas where they do not posses the domain knowledge or can sort through the amount of information in a timely manner. 

Make a Simple Decision:

Not only can cognitive augment knowledge and show alternatives for simple decisions, it can make a decision for humans and explain why it did so. In fact it can explain why other alternatives aren't as good as the one selected by a cog or a group of collaborating cogs. 

Make a Interconnected Decision:

Cognitive is really great at making a decision where there are a number of related decisions driven by shifting goals, priorities and weighting. This is a more difficult task for humans under fire and cognitive shines in this arena of thinking and making decisions. Cognitive can do this while providing the kind of transparency that shows trace ability and supports audit ability. 

Advise on an Action:

Cognitive can also suggest actions for humans to authorize and explain why and support various "what if" approaches. This can help build the confidence of the person authorizing actions. 

Decide and Take Action:

Cognitive not only can decide, it can act. This is where all the scary stories start emerging, but proper boundaries and rules will create the proper controls. This is especially necessary when multiple autonomous cogs are acting in concert in areas not encountered or anticipated. We have seen lots of success with robotics / IoT and I expect to see much more in the white collar arenas as well. 

Net; Net:

When it comes to taking in large volumes of data, information and sensory inputs in near real  time and answer questions, explain the best alternative and explain why other alternatives aren't as good, it's hard to ignore cognitive computing.