As often happens with experience, answers can be found hiding in plain sight, hiding from those too close to the issue, and obvious to those with a fresh but experienced perspective. A fresh perspective is what I’m writing about today.
Working in the technology industry, you tend to view things a little differently. Process and structure form the basis of understanding, employing neatly organized, methodical architecture and syntax, optimized to produce reliable results. Certainly in practice it can be as simple as it can be complex, and I suspect that is an inherent characteristic of creating something new.
But what about something that isn’t new? Something that we are comfortable using every day, something so familiar we share it with family, colleagues and even complete strangers. Despite a lifetime of use and experience we employ a blunt object approach in an attempt to make it useful.
Time’s up Ron. What are you talking about? I’m talking about language. I’m talking about cognitive analytics.
Business, and to an increasing extent government, both desire to better understand its one-to-many relationship with customers and citizens. Both want to leverage the collective asset that is the documented knowledge of its staff, operations and activities as well as services, search, e-learning, analysis, et al, all which is further exacerbated by our increasing electronic communication.
What do we do with all of this language of which we are so familiar but whose volume is becoming overwhelming? We abstract it of course.
More often than not, I see I mathematical approach—hidden under the guise of contemporary technology—being used to address any and all requirements. I frequently hear “well, I know it’s not what we wanted, but it’s better than what we had,” where the limitation is not the technologist but the technology. This unacceptable compromise comes with significant unintended consequences and cost.
Recently, I have been privileged to present these very concepts to a number of CIOs and accomplished enterprise architects. Having worked closely with this level of experience and sophistication for many years, I have come to expect the kind of detailed and probing questions relevant to the large enterprise. But that’s not what happened.
I’m not sure who was more surprised, those who prefaced our discussion by stating that “this domain, while under evaluation, is not likely to be attended to for many months”, or me. It seems that we have become conditioned to the notion of acceptable challenge and complexity and endeavor to attribute a familiar designation.
For example, I am often asked, “so you’re using lists, right?” No, I reply, no lists. I was recently surrounded by a roomful of PhDs who were absolutely certain that we “must be using math and algorithms to make this work”. Nope, we’re not. Our technology understands the meaning of words in a way that is similar to how you and I do. Hence, cognitive analytics.
A Wikipedia entry states: “Proficient reading is equally dependent on two critical skills: the ability to understand the language in which the text is written, and the ability to recognize and process printed text”.
It’s just that simple. Understand the words by their meaning, understand what words mean when they’re together, understand the emotion conveyed by the meaning, and finally, understand what is important in the meaning. To understand the language, you must first understand the words! Understanding, cognition, meaning: these concepts are at the heart of cognitive analytics.
Maybe it’s just that it seems too easy, or it’s a result of how some of the large vendors have trained us to accept known limitations and difficulties without pushing further. Sometimes, the answer may seem too simple or obvious to be credible. Which is why a fresh perspective can make all the difference in the world.
