Ten stitches, an ER visit, and a whole bunch o’ blood, but no cracked skull…proof that my head is thick.  (No wonder why my PhD took 10 years.)  And no, Mom, none of this has anything to do with the motorcycle.

I wish I had a better story to tell.  Something like hitting my head while pulling 8 g’s during barrel rolls in my experimental stunt airplane or getting gashed by a wayward spear gun shot while diving off the Great Barrier Reef.   No, no…I just smacked myself with one of these.

That’d be a post driver.  A tube with a 20 pound weight at one end that you let slam down on fence posts to drive them into the ground.  Last Thursday evening, I was working on the Chicken Incarceration Project (a story for another time) when I violated the first commandment of tool usage.  I didn’t let the tool do the work for me.  I was on a tough patch of ground, and I was putting some serious body english on that driver…really slamming it home.  Just as I started to pull it down for another whang, it caught on the top of the post and transferred all of its energy into a 90 degree rotation right on my head.  Yah, I know.  Idiotic.

What possible relevance does this have to Attensity and the world of NLP?  It reminded me of a tenet we try to live by here:  surround yourself with people who are smarter than you.  As last week demonstrated, that’s not so hard for me to do.  We practice that notion because we all want to be challenged and excited by our work environment, and ultimately that boils down to who you work with, not what you do.  Or at least I think so.  But hiring people smarter than yourself has solid business value when it comes to innovation.  Here’s a case in point.

Very early on at Attensity, we built our core engine.  It’s a chunk of highly efficient C++ code that effectively diagrams sentences.  It carves up language into things like subjects, direct objects, verb phrases, prepositional phrases, etc. Since most of us at the time had come from an academic research environment, our engine spit out its sentence diagrams in a very dense, information-heavy kind of way.   Like this:

Now while that sort of thing makes nut jobs like me giddy, it’s awfully taxing for everyone else to read.  Yet, it’s vital that we’re able to show our prospective customers the value of diagramming sentences.  Doing that with our engine’s raw output was hard.  Really hard.

Then one of our developers, who was never trained in the narrow field of Natural Language Processing, pondered the issue and wondered if he could turn our output into something more digestable.  In a couple of days, he put together a prototype that did this:

Already, moving from a character-based output to a graphical one was huge improvement, but that wasn’t what really blew us away.  He made the prototype reprocess the sentence after every single keystroke.  The result was that you could see how the diagramming process happened, in real-time.  Not only did it illustrate how the core engine worked, but it showed off how fast it was.  (The engine’s speed remains a point of pride here.)  When he demonstrated it to us, we thought it was the coolest thing we’d ever seen.  Nearly ten years on, his idea has become a key component of configuration and demonstration tools.   And, of course, now it seems obvious, but it wasn’t at the time.

I believe that one of the reasons why this fellow came up with one of our top 10 innovations is that he did not come from the NLP world like most of us.  While the rest of us assumed that the way we had done it all along was the best way, he thought about the problem differently, and that was the genius.  It just goes to show that innovation can not be prescribed.  The best you can do is provide fertile ground, and creating a group of smart, creative, and hard-working people is the best way I know of how to do that.  By the way, this fellow is also a super person…one of those people who seem to always be in a good mood.  Wish I could be more like that.

I have to go now and lock our VP of Engineering in his office again.  He keeps posting pictures of my ER visit on the breakroom refrigerator.  A big thanks to my wife who sent him those pics.  Grumble.

It’s hard to succeed without a great product, but it’s also hard to succeed without creating a strong perception of your product. A press release at the right time to the right places can help ensure your brand is always on top of the dialog on the web.

Community Insights allows you to monitor the dialog about your competitors. Updating in near real time, you can see when your competitors are getting press, what the sentiment is, and where they are getting coverage. If your PR team is agile, they can quickly issue messaging that can prevent your competitors’ campaign from gaining traction.

We used Community Insights to measure the dialog about three hosting cloud providers: RackSpace cloud, Microsoft Azure, and Amazon EC2. Here is the coverage for the past 30 days:

If you track the red line, you can see that on August 25, RackSpace announced the tools for their cloud. This was a pretty big announcement as these tools only make using their cloud easier. On the 26th, Amazon announced their virtual private cloud, which you can see by the spike in the blue line.

Would Rackspace’s announcement about its cloud tools have received more attention if Amazon had not launched their virtual private cloud the very next day? How does this affect people’s perception of what is available to people competitively shopping for cloud hosting?

Social media is a growing and useful part of the business landscape, not just “social” landscape.  A recent blog by Robert Scoble, touts the growing business value of Twitter.This widely viewed vignette on Youtube does a very entertaining job of underscoring what is happening in social media versus other media, from a more historical perspective.

I certainly realize that there is a great deal of skepticism that still remains regarding the value of social media for business.  On the other hand, I observe that this is a tsunami that one can ignore at our own peril.  The foundational basis for using social media in business is that business is all about customer conversations, and about these conversations no matter what the source of the conversation is.  So the immediate question that business must answer is “do I want to be involved in conversations with my customers?”  It seems to me the answer to that question is obvious.

The immediate follow-on to that question is “how to do that?”  It is our belief that people in companies who are charged with the responsibility for customers need a system of record to do so.  Increasingly, this responsibility is being centered in titles like Chief Customer Officer, Chief Customer Advocate, etc. If that is one’s responsibility, then how is that possible to get a handle on all the customer conversations that are occurring both within and outside your four walls?

Look around – other executives clearly have systems of record to support their responsibilities – the CFO, the VP Sales, the VP Manufacturing, and the list goes on.  If they have a system of record to support their responsibilities, is it not equally important for the Chief Customer Officer (CCO) to have one as well?  The alternative is to be a begger to these other organizations or to IT to “please provide me with a way to aggregate and consolidate all the forms of customer conversations that are happening with us and to us.”  As the old saying goes, ‘beggars cannot be choosers’, and that CCO will wind up as a victim and be ineffective.

If you agree with this point-of-view, then you owe it to yourself to check out the Attensity offerings.  We have uniquely created a system of record for these CCO’s and others with similar responsibilities in corporations and government.  Our system of record allows you to begin simply with something like surveys and proceed more expansively to other sources of customer conversations over time as experience and needs dictate.  While we can take in all kinds of sources of this data, we also have the ability to conduct signficant anaytics looking for trends and new issues, but further connecting to BI, data warehouses and other statistical systems for further work on predictive analytics.

My recommendation is not to take on such an important role as CCO without also having a view towards how you are going to fulfill those responsibilities.  In my opinion, that starts with a dialog about how technology can fulfill its role in making you successful – and that means a dialog about “what is my system of record?”  If you want to begin that dialog, we stand ready to share and help.

Hi all,

David Bean here.  At the risk of offending the marketing and PR side of Attensity, I’m going to tell you exactly how I feel about writing this blog.  Reluctant.  Uncomfortable.  Agitated.  Downright grumpy about the whole thing.  I wish I had a dental appointment right now.

I mean, you’ve gotta have a certain level of hubris to believe that anyone wants to read your blog / follow you on twitter / check your facebook status.  I’ve held out for nearly a year, but I’ve been brow-beaten, cajoled, and threatened enough.  (ooo…I’ve even beeninveigled.)  Alright, already…I’ll join the social media revolution…as a draftee.

Don’t get me wrong.  I get it.  I understand that social media are changing the way we  communicate, and Attensity and our many of our customers are using social media as sources in Voice of the Customer conversations.  But self-promotion just makes my skin crawl.  That’s not meant to sound like insincere modesty.  It’s just the way I feel about it.  And when I’m in one of those grass-is-greener-over-there moments, I think about being a park ranger.   Ah…just imagine:  you get to be outside, you can see the horizon and feel the breeze, you get to drive a pickup, you can take your black labrador named Moose with you everywhere, and if anyone gives you a hassle, you’re packing heat.  (Not that I’d know much about how to use one of those things.  I’m a lousy excuse for a Utahn, including the fact that I don’t carry a concealed weapon permit, but don’t get me started….)

Of course, I’m now contributing to one of my other issues with social media:  the vacuous nature of a lot of it.  Please, nobody take offense at that.  There’s plenty of good stuff out there, but there also is no built-in mechanism for something like peer review to ensure some level of quality.

Enough griping.  Until I find an opening at the Park Service, I promise to A) limit my self-publishing to topics that I think I might actually know something about (prepare to be amazed at my knowledge of how to unsuccessfully rebuild a Carter carburator for a 1950 Farmall C tractor) , and B) try to offer something of value to the sea of content.

- David

We are Champions – World Champions to be precise. Not in an athletics discipline but in computer cooking. This year, the 2nd International Computer Cooking Contest (CCC) took place in Seattle, Washington at the annual Case-Based Reasoning Conference. And the winner is: CookIIS. CookIIS is an intelligent Computer Cooking application built on top of  Empolis’ Information Access Suite and developed into a solution for cooks by the University of Hildesheim’s Intelligent Information Systems Group. So – what is this all about?

Technologies such as Natural Language Processing, Case-Based Reasoning and theSemantic Web are not self-explanatory. And, even worse, most of the applications related to these areas are not self-explanatory either. This drove my colleague Ralph Traphoener to the idea of creating an international contest around an everyday problem everybody understands to show the real word potential of these technologies easily: Everybody has some experience in cooking, no matter if they are good at it or not, hence, everybody knows what a recipe is and how complex it can be to execute well on.  It’s for turkey but you have got only chicken. No problem, adapt.

So we thought if we can create a Computer Cooking Application that shows the same behavior, then we can communicate the benefits of our technologies easily to a broader public. Finally, doing this as an international contest is fun. The International Computer Cooking Contest was born. In robotics such events already have a tradition, namely theRoboCup. Its vision is to beat the human FIFA World Champion in football/soccer by the year of 2050 with a team of humanoid robots.

The RoboCup is the role model for the International Computer Cooking Contest. The first contest (watch it on YouTube) was organized last year at the University of Trier, Germany and now the second event took place in Seattle. You can find the exact rules and terms on the CCC web site (CCC).

The basic idea is as follows: The participants receive a set of cooking recipes in advance of the contest. The recipes have semi-structured content, i.e. the ingredients list, and unstructured content, i.e. the preparation instructions. Now the teams have 3 months time to build their cooking applications. At the contest additional recipes may be supplied and several different types of tasks must be tackled on-site in front of the audience and the jury. The tasks are not known in advance. To give an example: “Create a three course menu consisting of an Asian soup, chicken for the main course. Do not use garlic. Low fat please.” Additionally, a list of the available ingredients is given. Watch it on YouTube.

The contestants propose recipes that are live evaluated by a jury.  A real chef is part of this jury and evaluates the taste of the solutions. Taste is the most important criterion for cooking results. Ever thought about the difference between lemon and lime? In the 2008 contest there were no lemons available but limes and oranges. One of the contestants proposed a recipe where it replaced a twist of lemon by lime juice. Chef Wolfgang Becker’sreaction was clear-cut: No go! The taste of lime is too intensive – it would destroy the whole dish. Orange would have been the right choice, i.e. also it is not the exact same taste as lemon,  but it fits the purpose of adding a fresh note to the dish. The lesson I learned: Cooking is not trivial. It requires knowledge about food and its properties.

Here is a quote from the CookIIS site that pins it down: “Something that seems to be simple and that many people manage easily turns out to be the cooperation of numerous complex fields of knowledge: When is a dish an appetizer? What makes a dessert Asian? Which are the chemical and flavorful functions an ingredient of a dish might have? What makes different ingredients similar or even exchangeable? When do two dishes harmonize and what makes a dish finally taste really good? Do all people like the same taste? The answers of these questions are based upon unconscious or only partly explicitly ascertainable experiential knowledge, otherwise they can only be known by trying out.”

The final and winning task that CookIIS solved best in 2009 was “Pizza with leek.” Among the given 1500 recipes there is (of course) none that directly fits. The system proposes a “No Meat Bean Burn Pizza” and it gives the hint to adapt the recipe by replacing the onions with the leek. Give it a try yourself.

The first system that performed this kind of task was Kristian J. Hammonds’ CHEF in 1986. He used this subject matter domain for studying cas-based planning models. We also have a sample application called Smartcooking for recipe retrieval that comes as standard withe:IAS. However, what can be seen at the Computer Cooking Contests is even more powerful, including text understanding, and tricky adaptation strategies, e.g. take a look atJaDaCook or Taaable, two of last year’s participants.

These cooking examples really show how computers can turn unstructured data into actionable information. They create awareness for the potential of really intelligent information systems based on these technologies. Now it is up to all of us and our creativity: If it can be done for cooking, then what else can be done? Keep an eye on theCCC web site -  as the winning recipe 2009 “Pizza with leek” already shows – next year’s dishes will be served in Italy.