Texas Wireless Summit: Big Brains, Big Data, and the Big Elephant in the Room
October 21, 2013 Leave a comment
Having never been to the Texas Wireless Summit (TWS13) before, my first clue of what I was in for should have been the venue’s name – The Applied Computational Engineering and Sciences Building at the University of Texas. Then, walking into a talk being given by Dr. Ashok Srivastava, Verizon’s Chief Data Scientist, the first words I heard were:
“…Taking vast amounts of machine learning to affect societal change.”
Great. Cue the proverbial pigs that fly and pie in the sky.
However, no sooner did I sit down and then talk quickly moved on from the ‘rise of the machines’ to delving into what ‘big data’ means to the mobile operator beyond implications that are merely advertising-based. From Verizon’s point of view (and, in fairness, most every other operator), customer churn and data monetization are two of the main business outcomes that big data is tasked with solving. Targeted advertising can be used to help with data monetization, but quite frankly, the technologists in the room at TWS13 were relatively unconcerned with the rudimentary activities of marketing hacks. They wanted to talk about, and hear about, how their life’s work is going to enable more than ads.
In that regard, I was pleased to find that this hack (blogger, analyst, consultant – take your pick) was singing from the same sheet of music. I’m kind of bored with the idea of analytics as the engine for bothersome push messages as I walk through the mall. There, I said it. If there’s going to be a use for big data analytics that justifies the billions being invested in technology development, then there have to be better commercial applications than just targeted advertising.
Now, at the end of the day, things did end up getting weird. After all, we were in Austin. The consensus of the room was that the next big thing in analytics, as taken from truly cutting edge disciplines such as aerospace and applied to telecom, is machines using big data to make decisions that humans are incapable of making at the speed required to be truly considered ‘real time.’ And when we get there, the applications are many. Just of the few more applicable, or interesting, include:
- Modeling when equipment will fail before it fails (think better quality of experience and even preemptive CapEx modeling);
- Modeling when patients’ organs will fail before they fail (think eHealth);
- Enabling SON to change network configurations that improve customer satisfaction at a granular level at speeds which humans cannot replicate (think, Amdocs’ acquisition of Actix, among a host of similar developments);
- Analyzing user data to devise service definitions on the fly for customer demographics that are difficult to predict (think teenagers with mobile plans akin to Google’s ‘Continuous Beta’ concept).
So, contrary to Dr. Srivastava’s commercial reality teaser, his main point – and the point(s) of most of his colleagues that followed – is that in order to realize the full potential of big data, the decisions and the resulting actions taken need to be fully automated. I must say, I agree (though this vision will clearly take years to realize).
Of course, the nagging question that remains to be answered was not asked by a journalist or analyst, but rather murmured among the intelligentsia assembled in the room who are leading the headlong sprint towards automation. “What happens to our brains when machines start making most of our decisions for us?”