I’ll Take Deep Learning for $2000, Alex



My smartphone made a mistake yesterday. I wanted to text, “I hear ya!” to a friend, which the autocorrect software changed to “I hear us.” I’m not sure how it decided that’s what I wanted
to say but I am sure it didn’t understand the context of the message. That, and a conversation with a friend, Richard Leavitt, CMO at AlchemyAPI, got me thinking about how the evolving ability to analyze unstructured data is both mind blowing and still imprecise, and how it will impact digital marketing, in particular content marketing and contextual advertising.

 

First, a dirty secret confession: I love Jeopardy!. I was glued to my couch during the
tournament between IBM’s Watson and past champions two years ago. So, it caught my
attention when IBM announced that an even more powerful Watson is now available via the Internet. Moving Watson from the Jeopardy! set to the cloud rewrites the economics of big data analysis, making huge levels of computational power available to just about anyone. It also sends a shot across the bow of the cloud services behemoth named Amazon Web Services, but that’s fodder for another post.

 

AlchemyAPI and Wayin are two Colorado companies that have recognized the potential to monetize Watson-esque artificial intelligence. According to a whitepaper from Alchemy, “Experts estimate that Fortune 500 companies are losing $12 billion per year in value because they are not exploiting unstructured data such as text.” Wayin’s emphasis is on social digital
behavior; their value proposition casts the company’s solution as “a social microsite assembly system.”


Readily available deep learning capacity offers researchers data analysis that enables not only the ability to accelerate such things as drug R&D and expand deep space exploration but to assist with decisions that could literally change humanity. But what you’ll see first is better content delivery, more accurate Google results and superior contextual ad placement. The reason for that it fairly obvious: the World Wide Web, which also encompasses the Socialverse, is where the data is most readily available and the future of commerce. Fact is, unstructured data in the form of documents, Tweets and video is the biggest mine of information on the Web.

 

What’s important to keep in mind is that, when it comes to marketing, there are advantages on the horizon for both sellers and buyers. For those hawking their products and services, it’s
manifested in more rapid and higher ROI for your assets, because there’s higher probability that they will be seen by the right people, at the right time. For you as a consumer, it means you should be getting hit with information that
actually is of value, whether unsolicited or requested. Everyone wins because digital behavior is being interpreted with greater nuance and sophistication, and more information is fueling the algorithmic ballet.

 

As for me, I’m looking forward to my smartphone getting even smarter, so when I tell it to say “I hear ya,” the person on the other end isn't confused. And, maybe, just maybe, Jeopardy! will host a tournament between the biggest deep learning services providers. That would be really smart.

I’ve Been Hacked!


I received a call from a woman named Andrea on Saturday morning in which she informed me that someone in New Hampshire had just tried to buy $211 from a grocery store, using a debit card number associated to SnydertheWriter LLC. Had I authorized the purchase? No. Was it on the card ending in XXXX? Yes. Was I in possession of the card? Yes, it’s in my hand, I said. 

 

They blocked the transaction, cancelled the card and notified my local bank. So, all’s well that ends well, right?

 

Maybe. In the “past lives” of my career, I’ve often written about how security impacts enterprises. And I’ve had the opportunity to exchange thoughts with some highly skilled security gurus while watching Internet traffic flow on a network map. Three areas in particular “resonate” with me, now that the experience of being hacked has transformed from the realm of supposition to reality.


Visibility

I called the bank and the credit card company on Monday, hoping that I might be able to learn where the security breech occurred. Sure, I understood that the fraudulent transaction happened in New Hampshire – and for the record, I’m in Longmont, CO, which is just about 2,000 miles away. What I wanted to know is where the data itself had been accessed and copied. Neither could tell me. The bank representative said “they don’t tell us that.” The credit card
representative said, “We can’t pinpoint the point of compromise.” 

 

So much for visibility; at least if what I was told is true. And I’ve used that card less than 10 times since I opened the account. The point I’m making is that if the good guys are going to at least compete on a level playing field with the bad guys, we need to be able to identify the point of compromise and shut that down. 

 

Frequency

First, let me say that I understand that credit card reps are trained professionals
who know how to use an even voice to keep down the blood pressure of the person they’re giving bad news. But gauging the calm in Andrea’s voice, I couldn’t help but think that this was as routine as ordering coffee at Starbucks for her. 

 

Obviously, accounts are being hacked all day long, all over the world. Have we reached the point where we should just expect security to be ineffective? Has security gone the way of privacy – just “get over it” that you don’t have any?

 

Impact

Freezing my debit card for at least two weeks has put on hold the purchase of a desk that I can work on while standing (about $200) and the tablet, monitor and
wireless mouse and keyboard I would have furnished it with (about $1,500
total). That’s $1,700 my small business is not putting into the economy, at least for the time being. Multiply that by who-knows-how-many businesses and
individuals who are in the same state of limbo. My bet is that the impact quite
rapidly starts to become significant.

 

Looking forward, I’m really hoping that this is a one-time event. On the one hand, it’s not a big problem. But on the other hand, IT IS a problem, especially if it happens again. One that we apparently have no control over. And it's a lucrative unprosecuted criminal 
activity for at least one person on the East Coast. 

 

Have you been hacked? What was the impact to you or your business?

Content development

 

Let's discuss!

Human Analytics: The Real Benefit of Big Data


The marriage of mind-boggling computer processing capabilities with oceans of
unstructured content is something that I believe heralds the next era of big data analysis -- human analytics. And the most significant benefit will be to enabling better quality of life in more places around the globe. Heck, it may even lead to better quality of life inside the walls of enterprise.

 

First, let’s get money discussed and out of the way. With natural language processing and deep learning tools in the hands of enterprise marketers, content and sentiment analysis will surely lead to increased revenue for companies smart enough to recognize that the more they understand what their customer care about,
the more successful they’ll be. My hat’s off to them! Making money is in fact a good thing and NLP will help forward-thinking organizations reap the rewards of using it.

 

Here’s where the human analytics  gets interesting. The brain is still superior to the
computer in some ways: nuance, subtlety, double-entendre, and originality are a few things that come to mind for me, pun intended. But what computers bring to the table is the sheer horsepower to serve up what you could not have known. For example, IBM’s Watson “can read all of the world's medical journals in less time than it takes a physician to drink a cup of coffee.” 
hope you stray from this post to read the article I’ve linked, but to summarize: After training Watson on lung cancer and feeding it daily with the latest relevant information published on the Web, the computer helped the physician diagnose and evaluate the course of treatment for the patient. The great thing is that Watson, and other Watson-esque NLP technologies, acts as an assistant, capable of pointing out what you might have missed, and is up to date in about as real-time as is possible on a given subject. "It’s like being able to take a knowledge

worker – cancer specialist, nurse, bond trader, portfolio manager, whatever – and equip that person with the best knowledge, and have it available at their fingertips."*

What’s also important is being able to share the results, so the patient on the other side of the world benefits from what’s learned.