Alleged Semantic Tips Spot on for Freshman Comp Students
August 5, 2015
I find the amount of attention given to semantic search as it applies to search engine optimization a fascinating development. “Semantic”, like Big Data, is fast becoming meaningless. The root of semantic is the Greek word for significant.
The application of the word semantic to information search and retrieval is a bit less straightforward. Toss in the concept of “search” and “content processing” and the output is an an information smoothie with big chunks of tough to identify systems and methods; for example:
- Methods to discern user intent
- Methods to figure out the context of an ambiguous element
- Programmatic data inserted into a content object which makes sense to a content processing system set up to recognize these instructions
- Systems which use pre-compiled look up tables or programmatic methods to figure out which words go together (White House or white house) or which alias goes with which person of interest
- Systems which attempt to “make sense” of content objects which signify some other information such as “Harrod’s teddy bear” as a token for an illegal substance
- Systems which deal with multi lingual corpuses
- Malformed Web accessible content which is supposed to comply with the W3C standards for semantic “stuff”.
You get the idea. Semantic drags in a number of interesting systems and methods. Many of these are complex and evolving as innovators try to deal with lousy precision and recall which is the norm for many “semantic” methods.
Now navigate to “Semantic Search Strategies That Work.” I would suggest that the tips in this write up apply to a person in an introductory college writing class. Here they are:
- “Forget about content as a daily grind.” Now that is music to a freshman’s ears. The silly notion that many professional writers have is that writing is something one must do every day and pursue with discipline. Nah, for real semantic search, take it easy. Chillax.
- “Concentrate on quality.” Now this is an interesting point. Google calculates quality based on a number of factors. The idea that a person who writes a high quality post and benefit from that effort is intriguing. In my experience, many excellent write ups get absolutely zero attention. These are usually write ups that address topics far from the pop music, Netflix, and Donald Trump scene. Here’s an example: Alon Halevy, et al, “Biperpedia: An Ontology for Search Applications.” This is a high quality paper, and I doubt that SEO mavens can match the effort which went into this 12 page write up. The write up deals with semantic issues, by the way.
- “When you write show who you are.” Not so fast. With the data lapses at various government agencies, health insurers, and corporate entities, content generated for the Web may require some thought, grooming, and vetting. How many SEO wizards want me to know about their behaviors and thoughts beyond their asserted expertise in fooling Google to rank an irrelevant site high in a query results list? How many SEO experts want the world to know that Google dropped a site in its rankings due to SEO missteps? What SEO expert wants a system to know what the person did prior to becoming an SEO expert? What about those secret actions like hunting lions in Africa or a dust up at a local watering hole? Think about this “who you are” stuff. Think carefully.
- “Focus on your prospects.” Ah, the bias is explicit. The motivating factor is that one writes to sell consulting work. Wrong. My hunch is that Dr. Halevy writes because he is curious and has colleagues with whom to collaborate in order to advance a particular area of inquiry. Halevy already sold a company to Google and, I assume, could sit at home and do volunteer SEO work. So far, he has resisted the siren song of easy money via baloney expertise.
- “Spend time on engagement.” I think this means attend conferences, post to social media, and hang out at watering holes without being captured in an on looker’s mobile phone picture.
Snake oil is available, gentle reader. Use with caution because it can damaged certain cognitive functions while emptying one’s bank account.
Stephen E Arnold, August 5, 2015
Dr. Watson: Concerned about Flabbiness and Sugar
August 5, 2015
The IBM PR attack continues. Today’s installment pits Watson (IBM’s Jeopardy winning, post production blind, Lucene based smart software) against flab and short chain soluble carbohydrates. Think diabetes or worse a visit to the dentist.
Navigate to “Dr. Watson: IBM Plans to Use big Data to Manage Diabetes and Obesity.” The story is not new. Once again IBM is reporting a “team up” deal. I wish the stories about Watson would talk about landing very large contracts with major government entities or Fortune 100 firms. I cannot get excited about old fashioned data mining applications. Sorry. Call me jaded.
The write up states without one whit of skepticism:
This new partnership marks a substantial leap into the healthcare sector for IBM, with CVS joining the likes of Apple and Medtronic as partners of IBM’s growing data service,Watson Health. By partnering up with CVS, Watson will be able to analyze and learn from “an unprecedented mix of health information sources”, including medical records, medical insurance claims and data from smart fitness devices.
I found the notion of the UK’s National Health Service hooking up with IBM an interesting one. Does the NHS have a functioning computer infrastructure? Has the promise of taxonomies delivered something useful to its intended users?
IBM might be able to help with systems. Will Watson remediate the NHS findability challenges? What will NHS pay to get Dr. Watson on the job? Has anyone involved in the Alphr (a former PC oriented outfit?) used Watson?
I don’t think much happens with these Watson stories than recycling what Watson’s team generates with rather amazing regularity.
Where are the billion dollar plus revenues? That is important to me.
Stephen E Arnold, August 5, 2015
Humans Screw Up the Self-Driving Car Again
August 5, 2015
Google really, really wants its self-driving cars to be approved for consumer usage. While the cars have been tested on actual roads, they have also been accompanied by car accidents. The Inquirer posted the article, “Latest Self-Driving Car Crash Injures Three Google Employees” about how the public might not be ready for self-driving vehicles. Google, not surprisingly, blames the crash on humans.
Google has been testing self-driving cars for over six years and there have been a total of fourteen accidents involving the vehicles. The most recent accident is the first that resulted in injuries. Three Google employees were using the self-driving vehicle during Mountain View, California rush hour traffic on July 1. When the accident occurred, each of the three employees were treated for whiplash. Google says that its car was not at fault and a distracted driver was at caused the accident, which is also the reason for the other accidents.
While Google is upset, the accidents have not hindered their plans, they have motivated them to push forward. Google explained that:
“ ‘The most recent collision, during the evening rush hour on 1 July, is a perfect example. The light was green, but traffic was backed up on the far side, so three cars, including ours, braked and came to a stop so as not to get stuck in the middle of the intersection. After we’d stopped, a car slammed into the back of us at 17 mph? ?and it hadn’t braked at all.’ ”
Google continues to insist that human error and inattention are ample reason to allow self-driving cars on the road. While it is hard to trust a machine with driving a weapon going 50 miles per hour, why do we trust people who have proven to be poor drivers with a license?
Whitney Grace, August 5, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
YouTube Wants You to Pay For…YouTube Content?
August 5, 2015
YouTube is free and that is one of the biggest draws for viewers. Viewers pull the plug on cable and instead watch TV and movies on the Internet or via streaming device. While YouTube might be free, video streaming services like Hulu, Netflix, and Amazon Prime offer network television for a fraction of the cable price. Google wants in on the streaming service game and it is already prepped with YouTube. Google’s only problem is that it does not have major TV networks signed up. Slash Gear explains in the article that “YouTube’s Upcoming Paid Service Hasn’t Signed Up TV Networks.” Cheaper access to network TV is one of main reasons that viewers sign up for a video streaming service, without them YouTube has a problem:
“What is most notable, however, is what is missing: TV networks. And according to sources, YouTube hasn’t at this point signed up any of those networks like NBC and Fox. Those networks would bring with them their popular shows, and those popular shows would bring in viewers. That doesn’t mean the networks will never be brought in — sources said there’s still time for them to get on board, as the rollout isn’t pegged for until later this year.”
Google is currently counting on YouTube stars to power the paid platform, which users will be able to watch ad free. Without network TV, a larger movie library, and other content, paying for YouTube probably will not have many takers. Why pay for already free videos, when all you have to do is watch a thirty-second ad?
Whitney Grace, August 5, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
The Race to Predict Began Years Ago: Journalism as Paleontology
August 4, 2015
I love reading the dead tree edition of the Wall Street Journal. This morning I learned that “Apple and Google Race to Predict What You Want.” The print story appears in the Business & Tech section on B1 and B6 for August 4, 2014. Note that the online version of the story has this title: “Apple and Google Know What You Want before You Do.” There is a difference for me between a “race” and “know.”
Nevertheless, the write up is interesting because of what is omitted. The story seems to fixate on mobile phone users and the notion of an assistant. The first thing I do with my mobile phone is find a way to disable this stuff. I dumped my test Microsoft phone because the stupid Cortana button was in a location which I inadvertently pressed. The Blackberry Classic is equally annoying, defaulting to a screen which takes three presses to escape. The iPhones and Android devices cannot understand my verbal instructions. Try looking up a Russian or Spanish name. Let me know how that works for you.
Now what’s omitted from the write up. Three points struck me as one which warranted a mention:
- Predictive methods are helping in reduce latency and unnecessary traffic (hence cost) between the user’s device and the service with the “answer”
- Advertisers benefit from predictive analytics. Figuring out that someone wants food opens the door to a special offer. Why not cue that up in advance?
- Predictive technology is not limited to a mobile applications. Google invested some bucks into an outfit called Recorded Future. What does Recorded Future do? Answer: Predictive analytics with a focus on time. The GOOG like Apple is mostly time blind.
Predictive methods are not brand, spanking new to those who have followed the antics of physicists since Einstein miracle year. For the WSJ and its canines, isn’t new whatever today seems bright and shiny.
Stephen E Arnold, August 4, 2015
IBM to Do the Apple Thing: Will Jeans and Black Turtlenecks Be the New IBM Fashion Trend
August 4, 2015
I don’t know if this article is accurate or a belated April Fool joke: “IBM to Purchase Up to 200,000 Macs Annually, with 50-75 Percent of Employees Ultimately Switching from Lenovo.”
My recollection is that IBM and Microsoft were once pals. Well, maybe sort of pals. There was that OS/2 thing, wasn’t there.
In the write up I learned:
A year after teaming up with Apple on an enterprise partnership to push iOS devices and apps for business users, IBM is moving forward with plans to rapidly move its own employees onto Apple’s platforms, MacRumors has learned. While IBM announced in an internal memo several months ago that it was planning to purchase up to 50,000 MacBooks for employees by the end of 2015, chief information officer Jeff Smith has revealed in a new internal video released to employees yesterday that he believes IBM could actually end up purchasing 150,000-200,000 Macs annually.
I am not sure if Lenovo is happy or sad to see this type of “news.” Years ago, I learned from IBM that it had “figure out Google.” I assume that this deal with Apple and the embrace of Apple hardware is part of that “figuring out.”
Stephen E Arnold, August 4, 2015
Coauthoring Documents in SharePoint to Save Time
August 4, 2015
SharePoint users are often looking for ways to save time and streamline the process of integration from other programs. Business Management Daily has devoted some attention to the topic with their article, “Co-authoring Documents in SharePoint and Office.” Read on for the full details of how to make the most of this feature.
The article begins:
“One of the best features of SharePoint 2010 and 2013 is the way it permits co-authoring. Co-authoring means more than one person is in a document, workbook or presentation at the same time editing different parts. It works differently in Word, Excel and PowerPoint . . . With Word 2013/SharePoint 2013, co-authors may edit either in Word Online (Word Web App) or the desktop version.”
SharePoint is a powerful but complicated solution that requires quite a bit of energy to maintain and use to the best of its ability. For those users and managers that are tasked with daily work in SharePoint, staying in touch with the latest tips and tricks is vital. Those users may benefit from Stephen E. Arnold’s Web site, ArnoldIT.com. A longtime leader in search, Arnold brings the latest SharePoint news together in one easy to digest news feed.
Emily Rae Aldridge, August 4, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Hire Watson As Your New Dietitian
August 4, 2015
IBM’s supercomputer Watson is being “trained” in various fields, such as healthcare, app creation, customer service relations, and creating brand new recipes. The applications for Watson are possibly endless. The supercomputer is combining its “skills” from healthcare and recipes by trying its hand at nutrition. Welltok invented the CaféWell Health Optimization Platform, a PaaS that creates individualized healthcare plans, and it implemented Watson’s big data capabilities to its Healthy Dining CaféWell personal concierge app. eWeek explains that “Welltok Takes IBM Watson Out To Dinner,” so it can offer clients personalized restaurant menu choices.
” ‘Optimal nutrition is one of the most significant factors in preventing and reversing the majority of our nation’s health conditions, like diabetes, overweight and obesity, heart disease and stroke and Alzheimer’s,’ said Anita Jones-Mueller, president of Healthy Dining, in a statement. ‘Since most Americans eat away from home an average of five times each week and it can be almost impossible to know what to order at restaurants to meet specific health needs, it is very important that wellness and condition management programs empower smart dining out choices. We applaud Welltok’s leadership in providing a new dimension to healthy restaurant dining through its groundbreaking CaféWell Concierge app.’”
Restaurant menus are very vague when it comes to nutritional information. When it comes to knowing if something is gluten-free, spicy, or a vegetarian option, the menu will state it, but all other information is missing. In order to find a restaurant’s nutritional information, you have to hit the Internet and conduct research. A new law passed will force restaurants to post calorie counts, but that will not include the amount of sugar, sodium, and other information. People have been making poor eating choices, partially due to the lack of information, if they know what they are eating they can improve their health. If Watson’s abilities can decrease the US’s waistline, it is for the better. The bigger challenge would be to get people to use the information.
Whitney Grace, August 4, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Facebook Wants You To Double Think About Using YouTube
August 4, 2015
Facebook does not like YouTube. Facebook wants to encourage users to upload their videos to its network, rather than posting them on YouTube. The Next Web shares how Facebook is trying to become major YouTube competition in “Facebook Throws Shade At YouTube When You Try To Paste A Link.” How is Facebook doing this? First, when a user tries to post a YouTube link, Facebook encourages users to upload to Facebook instead. Most users do not want to upload to Facebook, because it does not offer the same posting options as YouTube or does it?
Facebook has apparently upgraded how users can share their videos, including new features such as adding categories, sharing as an unlisted video, and disabling embedding. One drawback is that this could increase the amount of stolen videos. Some users might upload a stolen video, claim it as theirs, and reap the benefits. Facebook, however, does have user Audible Magic to catch a stolen copyrighted video. A direct quote from a Facebook representative said:
“ ‘For years we’ve used the Audible Magic system to help prevent unauthorized video content. We also have reporting tools in place to allow content owners to report potential copyright infringement, and upon receiving a valid notice we remove unauthorized content. We also suspend accounts of people with repeated IP violations when appropriate.’”
Thievery of original content is an important factor Facebook needs to work on if it wishes to rival YouTube. Popular YouTube celebrities and channels work hard to create original content and YouTube is a proven, marketable network. Facebook needs to offer competitive or better options to attract the big names, but for the average Facebook user uploading a video directly to Facebook is a desirable option.
Whitney Grace, August 4, 2015
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Poor IBM i2: 15 Year Old Company Makes Headlines in Fraud Detection and Big Blue Is Not Mentioned
August 3, 2015
Before IBM purchased i2 Ltd from an investment outfit, I did some work for Mike Hunter, one of the founders of i2 Ltd. i2 is not a household name. The fault lies not with i2’s technology; the fault lies at the feet of IBM.
A bit of history. Back in the 1990s, Hunter was working on an advanced degree in physics at Cambridge University. HIs undergraduate degree was from Manchester University. At about the same time, Michael Lynch, founder of Autonomy and DarkTrace, was a graduate of Cambridge and an early proponent of guided machine learning implemented in the Digital Reasoning Engine or DRE, an influential invention from Lynch’s pre Autonomy student research. Interesting product name: Digital Reasoning Engine. Lynch’s work was influential and triggered some me too approaches in the world of information access and content processing. Examples can be found in the original Fast Search & Transfer enterprise systems and in Recommind’s probabilistic approach, among others.
By 2001, i2 had placed its content processing and analytics systems in most of the NATO alliance countries. There were enough i2 Analyst Workbenches in Washington, DC to cause the Cambridge-based i2 to open an office in Arlington, Virginia.
i2 delivered in the mid 1990s, tools which allowed an analyst to identify people of interest, display relationships among these individuals, and drill down into underlying data to examine surveillance footage or look at text from documents (public and privileged).
IBM has i2 technology, and it also owns the Cybertap technology. The combination allows IBM to deploy for financial institutions a remarkable range of field proven, powerful tools. These tools are mature.
Due to the marketing expertise of IBM, a number of firms looked at what Hunter “invented” and concluded that there were whizzier ways to deliver certain functions. Palantir, for example, focused on Hollywood style visualization, Digital Reasoning emphasized entity extraction, and Haystax stressed insider threat functions. Today there are more than two dozen companies involved in what I call the Hunter-i2 market space.
Some of these have pushed in important new directions. Three examples of important innovators are: Diffeo, Recorded Future, and Terbium Labs. There are others which I can name, but I will not. You will have to wait until my new Dark Web study becomes available. (If you want to reserve a copy, send an email to benkent2020 at yahoo dot com. The book will run about 250 pages and cost about $100 when available as a PDF.)
The reason I mention i2 is because a recent Wall Street Journal article called “”Spy Tools Come to Wall Street” Print edition for August 3, 2015) and “Spy Software Gets a Second Life on Wall Street” did not. That’s not a surprise because the Murdoch property defines “news” in an interesting way.
The write up profiles a company called Digital Reasoning, which was founded in 2000 by a clever lad from the University of Virginia. I am confident of the academic excellence of the university because my son graduated from this fine institution too.
Digital Reasoning is one of the firms engaged in cognitive computing. I am not sure what this means, but I know IBM is pushing the concept for its fascinating Watson technology, which can create recipes and cure cancer. I am not sure about generating a profit, but that’s another issue associated with the cognitive computing “revolution.”
I learned:
In pitching prospective clients, Digital Reasoning often shows a demonstration of how its system respo9nded when it was fed 500,000 emails related to the Enron scandal made available by the Federal Energy Regulatory Commission. After being “taught” some key concepts about compliance, the Synthesys program identified dozens of suspicious emails in which participants were using language that suggested attempts to conceal or destroy information.
Interesting. I would suggest that the Digital Reasoning approach is 15 years old; that is, only marginally newer than the i2 system. Digital Reasoning lacks the functionality of Cybertap. Furthermore, companies like Diffeo, Recorded Future, and Terbium incorporate sophisticated predictive methods which operate in an environment of real time information flows. The idea is that looking at an archive is interesting and useful to an attorney or investigator looking backwards. However, the focus for many financial firms is on what is happening “now.”
The Wall Street Journal story reminds me of the third party descriptions of Autonomy’s mid 1990s technology. Those who fail to understand the quantity of content preparation and manual, subject matter expert effort required to obtain high value outputs are watching smoke, not investigating the fire.
For organizations looking for next generation technology which is and has been working for several years, one must push beyond the Palantir valuation and look to the value of innovative systems and methods.
For a starter, check out Diffeo, Recorded Future, and Terbium Labs. Please, push IBM to exert some effort to explain the i2-Cybertap capabilities. I tip my hat to the PR firm which may have synthesized some information for a story that is likely to make the investors’ hearts race this fine day.
Stephen E Arnold, August 3, 2015
-
- Subscribe to Beyond Search
Feature archive
News archive
-
