Watson: Will It Be Able to Make Major Government Applications More Intelligent?
June 30, 2015
I recently commented on the 25 percent problem rate in government software. You can find that Beyond Search item at this link. I can relate to companies who want to improve US government software. Go for it.
IBM has a plan which apparently ignores IBM Federal Systems (an outfit which creates, upgrades, and maintains some US government software). The approach focuses on a group of student from the University of Texas at Austin.
The write up states that Lauri Saft, director of the IBM Watson Ecosystem, has this view:
“You don’t program Watson, you teach it. We gave them [the students] the empty shell of Watson and said, ‘Go and come up with ideas you feel would be valuable.’”
The training method was one of Autonomy IDOL’s most important functions. The challenge which IDOL licensees faced, as I understand the system, is that the system can drift unless training and calibration are part of the routine maintenance cycle. For some licensees, the time and cost of the training and calibration were hurdles. Has Watson moved beyond Autonomy’s approach?
I assume the answer is, “Yes.” Therefore, the use of Watson to improve government software by making that software more intelligent should be a home run.
The students in Austin created a Watson app named CallScout. The focus is not on government software as I think of market opportunities. The students are working on a social service program in Texas. The application is customer support solution. That’s good.
My thought was that IBM Federal Systems would be using Watson’s remarkably broad spectrum of capabilities to address issues at the national level, maybe the regional level for Homeland Security or the EPA. I did not expect a local app for social services in Texas.
Perhaps the IBM Federal Systems Watson home run will be announced soon. The incubator with student thing is not likely to boost IBM top line revenues in the way a major US government Watson deal would. But PR is PR, big or small. IBM needs its Federal Systems’ unit to get the Watson revenue flowing. Time is a wastin’ because there are other outfits nosing into this potentially lucrative territory.
Stephen E Arnold, June 30, 2015
Keyword Search Is Not Productive. Who Says?
June 30, 2015
I noticed a flurry of tweets pointing to a diagram which maps out the Future of Search. You can view the diagram at http://bit.ly/1LznR1U or http://bit.ly/1dugOcR. Direct your attention to this assertion:
As amount of data grows, keyword search is becoming less productive.
Now look at what will replace keyword search:
- Social tagging
- Automatic semantic tagging
- Natural language search
- Intelligent agents
- Web scale reasoning.
The idea is that we will experience a progression through these “operations” or “functions.” The end point is “The Intelligent Web” and the Web scale reasoning approach to information access.
Interesting. But I am not completely comfortable with this analysis.
Let me highlight four observations and then leave you to your own sense of what the Web will become as the amount of data increases.
First, keyword search is a utility function, and it will become ubiquitous. It will not go away or be forgotten. Keyword search will just appear in more and more human machine interactions. Telling your automobile to call John is keyword search. Finding an email is often a matter of plugging a couple of words into the Gmail search box.
Second, more data does translate to programmers lacing together algorithms to deliver information to users. The idea is that a mobile device user will just “get” information. This is a practical response to the form factor, methods to reduce computational loads imposed by routine query processing, and the human desire for good enough information. The information just needs to be good enough which will work for most people. Do you want your child’s doctor to take automatic outputs if your child has cancer?
Third, for certain types of information access, the focus is shifting, as it should, from huge flows of data to chopping flows down into useful chunks. Governments archive intercepts because the computational demands of processing information in real time for large numbers of users who need real time access are an issue. As data volume grows, computing horsepower is laboring to keep pace. Short cuts are, therefore, important. But most of the short cuts require on having a question to answer. Guess what? Those short cuts are often keyword queries. The human may not be doing keyword searching, but the algorithms are.
Fourth, some types of information require both old fashioned Boolean keyword search and retrieval AND the manual, time consuming work of human specialists. In my experience, algorithms are useful, but there are subjects which require the old fashioned methods of querying, reading, researching, analyzing, and discussing. Most of these functions are keyword centric.
In short, keyword queries can be dismissed or dressed up in fancy jargon. I don’t think the method is going away too quickly. Charts and subjective curves are one thing. Real world information interaction is another.
Stephen E Arnold, June 30, 2015
The Google Cloud: Low Ceiling, Visibility Limited
June 30, 2015
I read “Google Cloud Platform: Google Execs Speak.” I highlighted one passage. In response to a question about recent Google cloud service price cuts, the Googler Brian Stevens said:
Our [pricing] is, to be honest, completely driven by measurable infrastructure improvements. So the numbers that you’re seeing aren’t even looking at the competition. They’re looking at the efficiencies. We actually can cost out all of our ongoing infrastructure for our platform, which we actually charge back to the group… We actually modeled those [costs]. We built our plans for next year. We have a set of goals around infrastructure efficiencies that we’re going to drive next year as well. Those [costs] are mapped right back into further and further discounts. So the model, for us, will continue.
I assume that Amazon will remain competitive with Google as both companies try to create value adding services. How low will Google cloud prices go? The suggestion that Google pays little attention to the actions of its competitors strikes me as interesting. I am sensitive to the words “honest” and “actually.”
Stephen E Arnold, June 30, 2015
Microsoft Puts the Cloud First with SharePoint Server 2016
June 30, 2015
Discussion of the cloud seems to push users into two camps: for and against. While hybrid is probably truly the way of the future, folks are still currently either of the “love it” or “hate it” variety. Redmond Magazine has provided good ongoing coverage of the upcoming SharePoint Server 2016 release, and their article, “Microsoft Taking a ‘Cloud First’ Approach with SharePoint 2016,” gives more details about what can be expected.
The article says:
“SharePoint Server 2016 will be a very cloud-inspired product when commercially released next year . . . Microsoft’s cloud services have been looming in the background of prior SharePoint Server releases . . . Office 365 cloud services have played a role since SharePoint Server 2013, and they will do so going forward with SharePoint Server 2016.”
One of the main promotional points of the new release is a promised “unified experience” for SharePoint users. While cloud skeptics still have reason to be cautious, the promised improvements may win them over. To stay up-to-date with the latest news regarding SharePoint, stayed tuned in to ArnoldIT.com and the dedicated SharePoint feed. Stephen E. Arnold is a longtime leader in search and his expertise comes in handy when trying to stay current without spending a lot of time doing independent research.
Emily Rae Aldridge, June 30, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Tumblr Has a GIF For You
June 30, 2015
Facebook recently enabled users to post GIF images on the social media platform. Reddit was in an uproar over the new GIF and celebrated by posting random moving images from celebrities making weird faces to the quintessential cute kitten. GIFs are an Internet phenomenon and are used by people to express their moods, opinions, or share their fandom. Another popular social medium platform, Tumblr, the microblogging site used to share photos, videos, quotes, and more, has added a GIF search, says PCMag in “Tumblr Adds New GIF Search Capabilities.”
The main point of Tumblr is the ability share content either a user creates or someone else creates. A user’s Tumblr page is a personal reflection of themselves and GIFs are one of the ultimate content pieces to share. Tumblr’s new search option for GIFs is very simple: a user picks the + button, clicks the GIF button, and then search for the GIF that suits your mood. A big thing on Tumblr is citing who created a piece and the new search option has that covered:
“Pick the GIF you want and it slinks right in, properly credited and everything,” the company said. “Whoever originally posted the GIF will be notified accordingly. On their dashboard, on their phone, all the regular places notifications go.”
GIFs are random bits of fun that litter the Internet and quickly achieve meme status. They are also easy to make, which appeals to people with vey little graphic background. They can make something creative and fun without much effort and now the can be easily found and shared on Tumblr.
Whitney Grace, June 30, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Webinar from BrightFunnel Ties Marketing to Revenue
June 30, 2015
The webinar on BrightFunnel Blog titled Campaign Attribution: Start Measuring True Marketing Impact (How-To Video) adds value to marketing efforts. BrightFunnel defines itself as platform for marketing analytics that works to join marketing more closely to revenue. The webinar is focused on the attribution application. The video poses three major questions that the application can answer about how pipeline and revenue are affected by marketing channels and specific campaigns, as well as how to gain better insight on the customer. The article overviews the webinar,
“Marketers care. We care a lot about what happens to all those leads we generate for sales. It can be hard to get a complete view of marketing impact when you’re limited to trusting that the right contacts, if any, are being added to opportunities! In this recording from our recent webinar, see how BrightFunnel solves key attribution problems by providing seamless visibility into multi-touch campaign attribution so you can accurately measure the impact you have on pipeline and revenue.”
BrightFunnel believes in an intuitive approach, claiming that three to four weeks has been plenty of time for their users to get set up and get to work with their product. They host a series of webinars that allows interested parties to ask direct questions and be answered live.
Chelsea Kerwin, June 30, 2014
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Alternative Search Engines: The Gray Lady Way
June 29, 2015
I read “Alternative Search Engines.” (Note: If you have to pay to read the article, visit a library and look for the story in the New York Times Magazine.) The process was painful. Distinctions which I find important were not part of the write up. The notion that some outfits actually index Web sites, and other outfits use Bing and Google search results without telling the user or the New York Times this cost cutting, half measure. Well, who cares? I don’t.
The write up asserts:
I was investigating the more practical, or just more traditional, alternatives to Google: Bing (owned by Microsoft), Yahoo (operated by Google back then and by Bing now), Info.com (an aggregator of Yahoo/Bing, Google and others) and newer sites like DuckDuckGo and IxQuick (which don’t track your search history), Gibiru and Unbubble (which don’t censor results) and Wolfram Alpha (which curates results). They were all too organized, too logical — the results were all the same, with only slight differences in the order of their presentation. It seemed to me that the Search Engine of Tomorow couldn’t be concerned with the best way to find what users were searching for, but with the best way to find what users didn’t even know they were searching for.
In case the Gray Lady has not figured out the real world, tomorrow means mobile devices. Mobile devices deliver filtered, personalized, swizzled for advertisers results. If you expect to run key word queries on the next iPhone or Android device, give that a whirl and let me know how that works out for you.
The crisis in search is that content is not available. Obtaining primary and certain secondary information is time consuming, difficult, and tedious. The reality of alternative search engines is that these are few and far between.
Do you trust Yandex.com or Yandex.ru? Do you know what the size of the Exalead search index is? What’s included and what’s omitted from Qwant, the search engine based on Pertimm (who?) which allegedly causes Eric Schmidt to suffer Qwant induced insomnia?
Nah. In Beyond Search, our view has been that the old fashioned, library type of research is a gone goose. The even older fashioned “talk to humans” and “do original research which conforms to the minimal guidelines reviewed in Statistics 101 classes” is just too Baby Boomerish.
With the Gray Lady explaining search, the demise of precision and recall, relevancy, editorial policies for inclusion in an index, and latency between information being available and inclusion in an index is history.
Stephen E Arnold, June 29, 2015
Loon Advances. Search Grounded.
June 29, 2015
I read “Google Talks Project Loon: 14 Different Prototypes, Leaks Solved by Using Fluffy Socks.” Loon is one of Google’s research projects. From what I can figure out, Google wants to provide Internet access to everyone, yep, categorical affirmative. Facebook has this idea too. The write up reports:
During Google I/O, he [Google wizard Astro Teller] said “We knew we had a lot to learn, but we misestimated how much we had to learn.” For example, the balloons are so large, they have to be stood on by engineers, and tests were carried out to see which socks caused fewer leaks. Fluffy ones, apparently, worked best. The research and perseverance has paid off. Project Loon’s balloons now stay in the air for six months, then steered around the world and positioned to within 500 yards of the intended target area. That’s way beyond the 100-day minimum flight time estimates, that Google says will make Project Loon a viable solution to provide Internet to the billions of people who cannot get it using traditional means.
One question: Will ads persist as long as a balloon remains aloft? Nah, longer. Search remains on the ground, with feet of clay.
Stephen E Arnold, June 29, 2015
Watson Gets Ink in the Bezos Newspaper
June 29, 2015
I read “The Human Upgrade: Watson’s Next Feat? Taking on Cancer.” The write up, which reminded me of an inclusion or sponsored content, states:
IBM’s computer brain is training alongside doctors to do what they can’t.
I am not exactly sure what this “do what they can’t” means, but the suggestion is that IBM Watson, which cannot generate significant revenue, can do something to ameliorate cancer.
Watson can put information together. Watson has written a cookbook. Then the write up startles with this statement:
But these feats were essentially gimmicks.
Gimmicks. No.
But now IBM Watson is not doing a gimmick:
The IBM program is one of several new aggressive health-care projects that aim to sift through the huge pools of data created by people’s records and daily routines and then identify patterns and connections to predict needs. It is a revolutionary approach to medicine and health care that is likely to have significant social, economic and political consequences.
Consequences like generating confusing or just plain wrong outputs? No way. This is IBM, the company which has declined in revenues for the last 36 months. IBM which spawns the often downright nasty commentary on Alliance@IBM. Impossible.
The write up summarizes some of the history of Watson. Omitted is the use of open source software to reduce certain costs. Left out is the mish mash of components which comprise Watson. Ignored are the human massage therapists essential to get information into Watson digestible form. Sidestepped are the computational requirements to build the index and then the tough problem of processing new or changed data in a timely manner so the outputs are not stale or outdated.
But Watson is evolving. Okay.
Like the Forrester professional who tosses around the word “revolution,” the write grabs this overinflated football.
Watson is a search and retrieval system. Watson adds layers upon layers of wrappers to impart value to outputs. IBM does not have a lock on smart software. What IBM does have is an ability to push the Watson story far and wide.
Does PR make a successful product? IBM obviously believes that marketing is the key to reversing Big Blue’s revenue challenges. IBM stakeholders may want to see some big contracts and zooming revenues. I know I do, and I am just a bystander deflecting PR flack in Harrod’s Creek.
Stephen E Arnold, June 29, 2015
Attivio ReachesTop 100 Status
June 29, 2015
The Data Dexterity Company announced the brand new Database Trends and Applications (DBTA) 100 and according to Yahoo Finance, Attivio is now on the list: “Attivio Named By Database Trends Applications To Its Prestigious Top 100 List.”
“We are pleased to be recognized by Database Trends and Applications as one of the most important firms in the data space; it further validates the type of feedback that our customers provide on a daily basis,” said Stephen Baker, CEO of Attivio. “As firms continue to be more reliant on maximizing their data to drive business-critical insights, we expect to play a critical role in driving this type of business innovation.”
Attivio joins the ranks of other companies that have made huge innovations in the data industry; they include EMC, Amazon, IBM, and more. Attivio is an industry leader in enterprise systems with its intelligence search platform. Attivio’s search platform enables users to make immediate insights with data visibility. Attivio has a well-known client use that encompasses such names as National Instruments, Nexen, GE, UBS, and Qualcomm. The company believes that there are many innovations to be made from all types, not just the type that is easily found in a database. Attivio uses its search platform to uncover insights in unstructured data that would otherwise be missed by other enterprise search platforms.
We have been following Attivio for many years and by having its name added to DBTA 100 proves it can perform well and deliver useful results. Enterprise search continues to be an important factor for enterprise systems, though people are often forgetting that today. Attivio’s addition to the DBTA 100 stresses that not everyone has forgotten.
Whitney Grace, June 29, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
-
- Subscribe to Beyond Search
Feature archive
News archive
-
