IBM Takes Step to Give Itself and Watson a Software Brain
April 11, 2015
Two observations to begin:
- MIT Technology Review recycles quite a bit of IBM marketing fluff. I now rely on the publication to keep me up to date on IBM’s floundering. Good stuff. Is there a flow of staff between IBM PR and the magazine?
- IBM continues to generate news which is designed to make the company seem to be the cutting edge outfit it once was. At least, the most recent development does not involve barbeque recipes with tamarind or revolutionizing medical treatment. That’s a welcome change, right, Watson?
I read “IBM Tests Mobile Computing Pioneer’s Controversial Brain Algorithms.” When I read the headline, I remembered that IBM has positioned its mainframes as mobile computing platforms. Well, I suppose doing the brain in math thing requires considerable computing horsepower. But will it work? Who knows? Certainly not the editors at dear old MIT Technology Review.
According to the write up:
IBM has established a research group to work on Numenta’s learning algorithms at its Almaden research lab in San Jose, California. The algorithms are being tested for tasks including interpreting satellite imagery, and the group is working on designs for computers that would implement Hawkins’s ideas in hardware. Hawkins says that around 100 people are working on the project, known internally as the Cortical Learning Center.
I quite like that “cortical” word. It sounds like important. Please, vocalize this statement with vocal fry.
IBM is in the midst of machine learning, artificial intelligence, and smart software. The write up points out that lots of companies are working hard to find ways to create smart software.
Numenta and probably the Cortical crowd want to find a middle way, a bit like the Great Compromiser except in digital form. The approach is not to make a system too simple. The approach is not to make a system too complex. IBM, like the children’s story, wants to make a brain in software just right.
Stay tuned, but if the software works, IBM’s revenues and profits and even its stock price may show some signs of life.
The life may be artificial, but life is life. Watson, isn’t life life? Where’s my Treo? Watson, Watson, do you know Mary Had a Little Lamp?
Stephen E Arnold, April 11, 2015
Twitter Plays Hard Ball or DataSift Knows the End Is in Sight
April 11, 2015
I read “Twitter Ends its Partnership with DataSift – Firehose Access Expires on August 13, 2015.” DataSift supports a number of competitive and other intelligence services with its authorized Twitter stream. The write up says:
DataSift’s customers will be able to access Twitter’s firehose of data as normal until August 13th, 2015. After that date all the customers will need to transition to other providers to receive Twitter data. This is an extremely disappointing result to us and the ecosystem of companies we have helped to build solutions around Twitter data.
I found this interesting. Plan now or lose that authorized firehose. Perhaps Twitter wants more money? On the other hand, maybe DataSift realizes that for some intelligence tasks, Facebook is where the money is. Twitter is a noise machine. Facebook, despite its flaws, is anchored in humans, but the noise is increasing. Some content processes become more tricky with each business twist and turn.
Stephen E Arnold, April 11, 2015
Search Is Simple: Factoid Question Answering Made Easy
April 10, 2015
I know that everyone is an expert when it comes to search. There are the Peter Principle managers at Fortune 100 companies who know so much about information retrieval. There are the former English majors who pontificate about next generation systems. There are marketers who have moved from Best Buy to the sylvan manses of faceted search and clustering.
But when one gets into the nitty gritty of figuring out how to identify information that answers a user’s question, the sunny days darkens, just like the shadows in a furrowed brow.
Navigate to “A Neural Network for Factoid Question Answering Over Paragraphs.” Download the technical paper at this link. What we have an interesting discussion of a method for identifying facts that appear in separate paragraphs. The approach makes use of a number of procedures, including a helpful vector space visualization to make evident the outcome of the procedures.
Now does the method work?
There is scant information about throughput, speed of processing, or what has to be done to handle multi-lingual content, blog type information, and short strings like content in WhatsApp.
One thing is known: Answering user questions is not yet akin to nuking a burrito in a microwave.
Stephen E Arnold, April 10, 2015
Elastic What: Stretching Understanding to the Snapping Point
April 10, 2015
I love Amazon. I love Elastic as a name for search. I hate confusion. Elasticsearch is now “Elastic.” I get it. But after I read “Amazon Launches New File Storage Service For EC2”, there may be some confusion between Amazon’s use of Elastic, various Amazon “elastic” services, and search. Is Amazon going to embrace the word “elastic” to describe its information retrieval system. Will this cause some confusion with the open source search vendor Elastic? I find it interesting that name confusion is an ever present issue in search. I have mentioned what happens when a company loses control of its name. Examples range from Thunderstone (a maker of search and search appliances) and the consumer software with the same name. Smartlogic (indexing software) is now facing encroachment from Smartlogic.io (consulting services). Brainware, now owned by Lexmark, lost control of its brand when distasteful videos appeared with the label Brainware. The brand was blasted with nasty bits. Where is the search oriented Brainware now? Retired I believe just as I am.
Little wonder some people have difficulty figuring out which vendor offers what software. Stretch your mind around the challenge of explaining that you want the Amazon elastic and the Elastic elastic. Vendors seem to operate without regard to the need to reduce signal mixing.
Stephen E Arnold, April 10, 2015
A Former Googler Reflects
April 10, 2015
After a year away from Google, blogger and former Googler Tim Bray (now at Amazon) reflects on what he does and does not miss about the company in his post, “Google + 1yr.” Anyone who follows his blog, ongoing, knows Bray has been outspoken about some of his problems with his former employer: First, he really dislikes “highly-overprivileged” Silicon Valley and its surrounds, where Google is based. Secondly, he found it unsettling to never communicate with the “actual customers paying the bills,” the advertisers.
What does Bray miss about Google? Their advanced bug tracking system tops the list, followed closely by the slick and efficient, highly collaborative internal apps deployment. He was also pretty keen on being paid partially in Google stock between 2010 and 2014. The food on campus is everything it’s cracked up to be, he admits, but as a remote worker, he rarely got to sample it.
It was a passage in Bray’s “neutral” section that most caught my eye, though. He writes:
“The number one popular gripe against Google is that they’re watching everything we do online and using it to monetize us. That one doesn’t bother me in the slightest. The services are free so someone’s gotta pay the rent, and that’s the advertisers.
“Are you worried about Google (or Facebook or Twitter or your telephone company or Microsoft or Amazon) misusing the data they collect? That’s perfectly reasonable. And it’s also a policy problem, nothing to do with technology; the solutions lie in the domains of politics and law.
“I’m actually pretty optimistic that existing legislation and common law might suffice to whack anyone who really went off the rails in this domain.
“Also, I have trouble getting exercised about it when we’re facing a wave of horrible, toxic, pervasive privacy attacks from abusive governments and actual criminals.”
Everything is relative, I suppose. Still, I think it understandable for non-insiders to remain a leery about these companies’ data habits. After all, the distinction between “abusive government” and businesses is not always so clear these days.
Cynthia Murrell, April 10, 2015
Stephen E Arnold, Publisher of CyberOSINT at www.xenky.com
Predicting Plot Holes Isn’t So Easy
April 10, 2015
According to The Paris Review’s blog post “Man In Hole II: Man In Deeper Hole” Mathew Jockers created an analysis tool to predict archetypal book plots:
A rough primer: Jockers uses a tool called “sentiment analysis” to gauge “the relationship between sentiment and plot shape in fiction”; algorithms assign every word in a novel a positive or negative emotional value, and in compiling these values he’s able to graph the shifts in a story’s narrative. A lot of negative words mean something bad is happening, a lot of positive words mean something good is happening. Ultimately, he derived six archetypal plot shapes.”
Academics, however, found some problems with Jockers’s tool, such as is it possible to assign all words an emotional variance and can all plots really take basic forms? The problem is that words are as nuanced as human emotion, perspectives change in an instant, and sentiments are subjective. How would the tool rate sarcasm?
All stories have been broken down into seven basic plots, so why can it not be possible to do the same for book plots? Jockers already identified six basic book plots and there are some who are curiously optimistic about his analysis tool. It does beg the question if will staunch author’s creativity or if it will make English professors derive even more subjective meaning from Ulysses?
Whitney Grace, April 10, 2015
Stephen E Arnold, Publisher of CyberOSINT at www.xenky.com
Enterprise Search and Marketers: Think Endpoint Computing
April 9, 2015
I have to hand it to the mid tier consultants. Just when I thought the baloney about enterprise search had begun to recede, I learned I was wrong. That puts me in my place.
Search is now “endpoint computing.” I know this because I received an email from the incubator-spawned X1 search company. I have tested X1 over the years, and I have come to think neutral thoughts about the company’s administrative options and its interface.
The method of communicating with me was a somewhat dry email that began with the salutation, “Hello.”
The email offered me a report by the ever fascinating Gartner Group. The point of the email is that X1 is a cool vendor. That’s nice. Curious I clicked on the link and was redirected to this page:
Okay, a lead generating system. I filled out the information and then I received another email. This one was a bit more serious.
The author, an earnest person named “Janice” wanted to speak with me to discuss my search requirement. Furthermore the person looks forward to speaking with me about “unified search and discovery for virtual, cloud, and hybrid environments.” X1 was founded in 2003and has experienced several management changes, which is common in the “unified search and discovery for virtual, cloud, and hybrid environments” market.
What makes X1 cool? To answer the question I had to read the Gartner Report, a task which I know is a chore.
The idea is that search is now endpoint computing. Okay. I guess. The report reassures me that the information in the report is not an “exhaustive list of vendors.” That’s good because in the report there are five companies mentioned:
- Login Consultants, a workspace consultant, but I don’t know what this term means
- Tanium, a company offering endpoint security and systems management, which strikes me as a consulting outfit
- X1, a search and retrieval vendor offering desktop search, eDiscovery, and enterprise search
- Kaviza (a where are they now company which puzzles me) a virtualized desktop outfit now owned by Citrix
- Framehawk (another where are they outfit), a company in the high definition user experience business (I have no idea what this means). Apparently Citrix does because Citrix also acquired Framehawk.
Quite an eclectic list. I remember when I worked at Ziff Communications in Manhattan. I listened to a group of editors working up a list of top trends over lunch. So much for methodology. The approach produced a somewhat eclectic list which was, in my opinion, of little value. The list was silly. But these were professionals. Who was I?
So the Gartner list is neither exhaustive nor coherent from my point of view.
What’s cool about X1 search as endpoint computing?
According to the mid tier consulting firms’ authors, X1 is cool because:
“Implementing VDI that provides a user experience that’s equal or superior to a distributed PC environment has been a huge challenge for organizations. While much of the innovation in the VDI space over the past few years has been focused on reducing cost and complexity, some vendors, like X1, have concentrated on removing barriers or exceptions that make VDI a compromise rather than a business enabler.” (page 3)
In the context of the firms profiled by Gartner’s “expert, the explanation of the X1 cool factor baffles me. I am not confused. I just don’t know what Gartner is trying to communicate.
I have several thoughts running through my head:
First, Gartner obviously has a financial model in place that makes it possible for the mid tier consulting firm to crank out analyses that seem to be authoritative. On closer inspection, the terminology and the information provided are not particularly useful. Does Gartner write these for free and allow the “cool” vendors to distribute these analyses for free? Why do I get a copy for free? Hmmm.
Second, there are obviously companies which value the Gartner endorsement even if it is not exactly clear what the message is. These companies—specifically X1—have seized upon the Gartner report as a way to generate leads and sales. I have no problem with that, but sending information that makes sense would appeal to me more than what I perceive as “information free” commentary.
Third, I continue to worry about the chance for meaningful discourse about the relative merits of information retrieval systems. The presentation of vendors in the context of buzzwords does little to convince me of the merits of X1 or the credibility of Gartner Group. I suppose that is why there are blue chip consulting firms and mid tier (azure chip) consulting firms. One good point: Unlike IDC’s Dave Schubmehl, the report was not $3,500 available on Amazon with my name slapped on as the “author.”
Score one for Gartner’s merrie band.
Stephen E Arnold, April 9, 2015
Twitter Search: Well, Sort Of
April 9, 2015
I read “Updating Trends on Mobile.” I am more interested in more detailed information about Twitter content, users, and tags. General purpose or massified outputs are of little utility in my little world.
I noted this passage:
We’ve been working to make content easier to find over the last several months in places like your home timeline – with recaps and Tweets from within your network – and through efforts like MagicRecs. We’ll continue to make improvements like these in the future.
If you navigate to the Twitter search page and enter a string like “enterprise search”, you will see variants of the term or phrase expressed as Twitter hash tags. The trends displayed were reflective of what Twitter’s log suggest is hot. Here’s an example:
How many of these trends do you recognize. I knew about iOS 8.3, Apple Watch, and not much else.
Queries for tweets remain a bit problematic for me.
Stephen E Arnold, April 9, 2015
Image Search Might Not Be Enough
April 9, 2015
Did you know there is a Google of image search? No, it is not the image option on the actual Google search engine. Rather it is Giphy aka the Google of GIFs (and a way to kill an hour) is stepping up into the world by buying other startups. TechCrunch reports that, “Giphy’s First Acquisition, Nutmeg, Is A Big Step Towards Mobile.”
Giphy has been interested in expanding its mobile search offerings and they recently acquired Nutmeg, a mobile GIF messaging app that makes it easier to send those fun moving pictures in a text message. Giphy founder Alex Chung and Nutmeg founder Julie Logan have discussed a partnership for the past year and after a recent $17 million round of funding by Giphy it felt like the right time.
“ ‘Nutmeg and Giphy share the same philosophy, but Julie brings a lot of expertise around what we’re doing from the mobile perspective, and that’s invaluable,’ said Chung. ‘The simplicity, the curation and the UX and the UI, drew us to Nutmeg.’ ”
GIFs are a universal Internet language with many of them transforming into memes and making the Reddit rounds. GIFs lucrative market due to their popularity and there is money to be made there.
Whitney Grace, April 9, 2015
Stephen E Arnold, Publisher of CyberOSINT at www.xenky.com
The Cost of a Click Through Bing Ads
April 9, 2015
Wow. As an outsider to the world of marketing, I find these figures rather astounding. MarketingProfs shares an infographic titled, “The 20 Most Expensive Bing Ads Keywords.” The data comes from a recent analysis by WordStream of 10 million English keywords, grouped into categories. Writer Vahe Habeshian tells us:
“WordStream analyzed some 10 million English keywords and grouped the them into categories to determine the most expensive types of keywords (see infographic, below).
“(Also see a similar analysis of the most expensive keywords in Google AdWords advertising from 2011.)
“The most expensive keyword on Bing Ads is ‘lawyer,’ which would cost advertisers seeking the top ad spot a whopping $109.21 per click. Not surprisingly, the top 5 keywords are related to the legal world, indicating how lucrative clients can be.”
Yes, almost $110 per click whether legitimate, a human error, or a robot script. That’s a lot of fruitless clicks. It seems irrational, but it must be working if companies keep spending the dough. Right?
The word in second place, “attorney,” comes to $101.77 per click, and “DUI” is a comparative bargain at $68.56. After the top five, law-related words, there are such valuable terms as “annuity,” “rehab,” and “exterminator.” See the infographic for more examples.
Cynthia Murrell, April 09, 2015
Stephen E Arnold, Publisher of CyberOSINT at www.xenky.com
-
- Subscribe to Beyond Search
Feature archive
News archive
-
