Figure 1: Cognitive computing spans across analytics landscape using all available data
Where to start? Like many emerging technological eras that some of us have witnessed in the past 25 years, cognitive computing
too needs to go through a ramp up period before it is widely adopted within the enterprises and across them.
It is only during this time that early adapters - market leaders and visionaries - would rip the benefits of the new technology by being the first to take advantage of it, taking the lion share of the value and most importantly influencing the way it shapes commodities in the future. However, gaining advantage of the new technological era over competitors requires businesses to acquire new skills among their professionals.
What skills are necessary to make this journey? To answer this question let’s take a look at a typical decision
processes as follows:
- Define the business problem using its unique domain, attributes and relationships.
- Translate the relationships between these attributes into a logical model that represents the business problem
- Connect the model to its relevant data in a form that is consumable by its solving algorithm
- Run the algorithm to solve the model
- Translate the result into a solution to the business problem.
To perform the above tasks, typically we need to have business analysts, who understand the business domain, mathematicians and/or statisticians who are able to build the logical models and design the algorithms to solve them, data analysts who know where the necessary data is stored, how to get them, make the necessary transformation and make them available to software engineers who know how to install, configure and connect all the software tools. The above description is perhaps an oversimplification of the skill requirements, however to understand where these skills would come from let’s take a couple of short steps back into how the science of problem solving has evolved.
Usually, we have two separate sets of disciplines thought at universities; these are the fields of Operations Research and
Computer Science that are heavily contributing to advancements in Decision Science, a branch of Management Science.
In the 70s and 80s, scientists from these fields were often working to invent new algorithms, or improve the existing ones
to solve better the same complex problem arising from management science without much visibility on each other work!
Most of this work was often carried out to overcome the limitations in available storage capacity, processing power and rapid access memory
which were limiting the size and complexity of the problems that could be solved in a rerasonable timeframe.
In the mid-90s however, thanks to a lot of pioneering joint work by scientists from both disciplines this started to change and new trends in science of complex problem solving began to emerge, see for example:
The advancements in algorithms and computer platforms for solving large scale complex problems has given businesses enough confidence to apply them in their daily business management and decision making process, see for example,
The above capabilities coupled with technology to obtain, analyse and process large volume of data has enabled companies to create insight that is consumable by decision makers be it man, machine or a combination of the two! Data Scientists (a newly created role in the marketplace after emergence of the big data) are key in delivering this. I have come across many different definitions for the role or skills that a Data Scientist must have, but the best that I have seen so far is by my favourite CDO @usamaf, who wrote:
"A Data Scientist is someone who Knows a lot more software engineering than Statisticians & Knows a lot more Statistics than software engineers"
Why Data Scientists are important in deploying cognitive computing? Cognitive computing as seen on http://www.ibm.com/cognitive/outthink/ is a journey, a journey that takes companies to the next level, a journey that involves planning and execution of all different steps along the way.
As an example, these steps could include but not limited to:
- Understanding the business long/mid-term goal and strategy.
- Aligning the strategy with cognitive capabilities to determine a road map for acquiring data and platforms necessary to achieve the goals.
- Map out man and machine engagement models, i.e. determines who learns from who, what? Who does what?
- Identify the business processes that directly benefit from digitization.
- And last but not least enable business innovation through cognition across all part of the business.
However, although this journey is different for every company, the vehicle is going to be the same. In my opinion Data Scientists are best positioned to drive this vehicle. They know the passengers and their destinations, they know the road and its turns and bumps, they know the vehicle and how it works, and most importantly they know what fuels this vehicle, Data!
Disclaimer: Unless specified, all thoughts and opinion expressed here are my own! @MoziHajian