In-Q-Tel and Algorithmia: The Widget Approach to Tasks
July 27, 2016
I noted “Algorithmia Lands In-Q-Tel Deal, Adds Deep Learning Capabilities.” Useful write up but one important facet of the deal was omitted from the news item. The trend in some government projects is to have a “open” set of software. The idea of running a query for a needed widget on an ecommerce site hath its charms. The goal is cost reduction and reducing the time required to modify an existing software system. This is the idea behind the DCGS (Distributed Common Ground System) widget approach. The issue I have with brokering algorithms is that most of the algorithms which look like rocket science are actually textbook examples or variations on well known themes. Search and content processing chugs along on 10 methods. When a novel solution becomes available like this insight into Carmichael numbers appears, years may pass before the method can be verified and inserted into the usable methods folder. Now those doing the searching have to know what an algorithm actually does and what settings are required for the numerical recipe to generate an edible croissant. That human knowledge thing is an issue.
Stephen E Arnold, July 27, 2016
X1 Hits a Familiar Marketing Chord
July 27, 2016
I had the job of reviewing some of the 100 eDiscovery systems on offer today. I noted an interesting semantic thrust in “Recent Court Decisions, Key Industry Report Reveal Broken eDiscovery Collection Processes.”
X1 states:
What is needed is an effective, scalable and systemized ESI collection process that makes enterprise eDiscovery collection much more feasible. More advanced enterprise class technology, such as X1 Distributed Discovery, can accomplish system-wide searches that are narrowly tailored to collect only potentially relevant information in a legally defensible manner. This process is better, faster and dramatically less expensive than other methods currently employed.
I highlighted the point that targeted search and collection can reduce the costs of certain legal processes. Recommind made, prior to its acquisition by OpenText, similar statements about cost reduction.
It strikes me that the number of 90 percent cost reduction begs for some hard data. Also, with so many eDiscovery firms competing for law firms’ business, is this once lucrative niche in a race toward Amazon- and Walmart-type pricing?
Perhaps there are several issues involved with the problem lawyers have when using eDiscovery systems? These may be:
- Unrealistic expectations for an eDiscovery system. Google does one thing; eDiscovery does another.
- Sticker shock. Despite the low cost of a license fee for a cloud service, there are other costs. These range from customer and engineering support to the need to pay for resources that cost money without doing much to alert the customer that a threshold has been crossed. Taxi meter and old school AT&T network pricing may dampen some eDiscovery licensees’ enthusiasm
- Federating content from multiple sources sounds wonderful. In my experience, what does one do about exception documents? What do eDiscovery systems “miss”?
As the financial pressure mounts on many law firms, cost controls are likely to rework how these companies do business. The impact on “search” vendors now in the eDiscovery sector may find themselves facing the type of financial realities which caused Recommind to reinvent its future as a company.
Stephen E Arnold, July 27, 2016
OpenText Buys CEM Platform from HP Inc.
July 27, 2016
When Hewlett Packard split up its business in 2015, consumer-printer firm HP Inc. was created; that entity got custody of HP’s CEM platform. Now we learn, from an article at TechCrunch, that “OpenText Acquires HP Customer Experience Content Management for $170 Million.” OpenText expects the deal to generate between $85 million and $95 million in its first year alone. Writer Ron Miller describes:
“The package of products sold to OpenText today come from the HP Engage line and includes HP TeamSite, a web content management tool left over from the purchase of Interwoven (which was actually bought by Autonomy before Autonomy was sold to HP), HP MediaBin, a digital asset management solution, HP Qfiniti, a workforce optimization solution for enterprise contact center management, as well as HP Explore, HP Aurasma, and HP Optimost.”
Some suspect HP was eager to unload this division from the time of the company’s split. Even if that is true, OpenText seems poised to make a lot from their investment; Miller cites the blog post of content-management consultant Tony Byrne:
“The most important thing to understand, though, is that as a vendor OpenText is a financial construct in search of a technology rationale. The company follows a ‘roll-up’ strategy: purchasing older tools for their maintenance revenue streams, streams which — while not always large — are almost always very profitable.”
It is true. In contrast to, say, Google’s method of trying nearly every idea conceived within their company and seeing what sticks, OpenText tends to be deliberate and calculated in their decisions. We are curious to see where this investment goes.
Based in Waterloo, Ontario, OpenText offers tools for enterprise information management, business process management, and customer experience management. Launched in 1991, the company now serves over 100,000 customers around the world. They are also hiring in several locations as of this writing.
Cynthia Murrell, July 27, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Salesforce Blackout
July 27, 2016
Salesforce.com is a cloud computing company with the majority of its profits coming from customer relationship management and acquiring commercial social networking apps. According to PC World, Salesforce recently had a blackout and the details were told in: “Salesforce Outage Continues In Some Parts Of The US.” In early May, Salesforce was down for over twelve hours due to a file integrity issue in the NA14 database.
The outage occurred in the morning with limited services restored later in the evening. Salesforce divides its customers into instances. The NA14 instance is located in North America as many of the customers who complained via Twitter are located in the US.
The exact details were:
“The database failure happened after “a successful site switch” of the NA14 instance “to resolve a service disruption that occurred between 00:47 to 02:39 UTC on May 10, 2016 due to a failure in the power distribution in the primary data center,” the company said. Later on Tuesday, Salesforce continued to report that users were still unable to access the service. It said it did not believe “at this point” that it would be able to repair the file integrity issue. Instead, it had shifted its focus to recovering from a prior backup, which had not been affected by the file integrity issues.”
It is to be expected that power outages like this would happen and they will reoccur in the future. Technology is only as reliable as the best circuit breaker and electricity flows. This is why it is recommended to back up your files in more than one place.
Whitney Grace, July 27, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Alphabet Google May Dominate AI for Developers
July 26, 2016
Advertising dominance is not enough. Tackling problems like solving death is not enough. Failing at augmented reality is not enough. Google wants to dominate artificial intelligence.
I know. I read “Google Sprints Ahead in AI Building Blocks, Leaving Rivals Wary.” Now quite a few folks have released to open source various types of smart software. But the Alphabet Google thing knows that smart software can cut electricity bills and generate less and less useful search results for me.
The write up takes a different approach. I learned:
But for some competitors, there’s a big downside to adopting Google’s standard. Using TensorFlow will help Google recruit more AI experts by training them on the same tool it uses internally, spotting their code, and hiring the best contributors. It could also let the search-engine provider exert outsize influence over the burgeoning AI ecosystem. If the internet giant dominates in this field, it could gain an advantage in the fast-growing cloud-computing business, turning the popularity of its software into real revenue. “It’s the next big area, and people are worried Google’s going to own the show,” said Ed Lazowska, a computer science professor at the University of Washington who has served on the technical advisory board of Microsoft Corp.’s research lab. “There is a network effect, and it’s a really excellent system.”
What’s the solution? Well, there is Facebook AI, IBM Watson, and a number of other outfits eager to make your software smart.
What’s not to like? Perhaps a silver bullet is speeding on its way to you?
Stephen E Arnold, July 26, 2016
Google and Its Algorithmic Methods Catch Attention in India
July 26, 2016
I read “Google in Legal Trouble in India for Including PM Modi in Top Criminals Search Results.” I have no way of knowing if the write up is accurate or the product of the great Internet news machine. I found the factoids amusing.
I highlighted this passage as fodder for the spirit of Jack Benny or Fred Allen:
The complainant, an advocate named Sushl Kumar Mishra, has argued that a Google search for the “top ten criminals of the world” showed Modi’s photograph in the search results. Mishra has said that while he wrote to Google to remove Modi’s name from the list, he got no response.
Customer service? In response to an email? Okay, this person is an optimist. As an aside, politicians are good actors. Right?
Google’s response was interesting as well:
The company had explained that the results were due to the British daily using an image of Modi with erroneous metadata and that the news articles did not actually link Modi to criminal activity.
Love that smart software.
Stephen E Arnold, July 26, 2016
Avast: Pirate Libraries
July 26, 2016
They are called “pirate libraries,” but one would be better-served envisioning Robin Hood than Blackbeard. Atlas Obscura takes a look at these floaters of scientific-journal copyrights in, “The Rise of Pirate Libraries.” These are not physical libraries, but virtual ones, where researchers and other curious folks can study articles otherwise accessible only through expensive scientific journal paywalls. Reporter Sarah Laskow writes:
“The creators of these repositories are a small group who try to keep a low profile, since distributing copyrighted material in this way is illegal. Many of them are academics. The largest pirate libraries have come from Russia’s cultural orbit, but the documents they collect are used by people around the world, in countries both wealthy and poor. Pirate libraries have become so popular that in 2015, Elsevier, one of the largest academic publishers in America, went to court to try to shut down two of the most popular, Sci-Hub and Library Genesis.
“These libraries, Elsevier alleged, cost the company millions of dollars in lost profits. But the people who run and support pirate libraries argue that they’re filling a market gap, providing access to information to researchers around the world who wouldn’t have the resources to obtain these materials any other way.”
The development of these illicit repositories traces back to Russia and its surrounds, where academics had a long history of secretly sharing documents under the repressive Soviet Union. In the 1990s, this tradition began to move online; one of the first pirate-library websites was Lib.Ru. Since then, illegally shared knowledge from more parts of the world has been made available, particularly from Western publishers and universities. Furthermore, the speed with which materials make it online has increased considerably.
Which is more worthy: protecting the stranglehold academic journals have managed to legally establish, and profit from, on research and other information? Or allowing people who possess great curiosity, but who lack deep pockets, to access the latest research? The scholarly pirates have made their choice.
Cynthia Murrell, July 26, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark Web meet up on July 26, 2016. Information is at this link: http://bit.ly/29tVKpx.
Azure Media Services Meets Microsoft Machine Learning
July 26, 2016
The article on TechCrunch titled Microsoft Is Bringing Automatic Video Summarization, Hyperlapse, OCR, and More to Azure Media Services discusses the machine learning enhancements coming soon to the cloud-based resources for video workflows. Media Services will be able to summarize videos, perhaps more honestly than the average theatrical trailer. Face detection is another perk. The article details,
“Microsoft is building face detection into these tools and the company is including its ability to detect people’s emotions (something the company’s Cognitive Services already do for still images). Using this, you could easily see how people reacted to a speech at an event, for example. If your keynote goes on for too long, you will probably see people’s emotions go from happiness to indifference and then to sadness and contempt.”
The face detection and feature-reading technology is also shown in CaptionBot, Microsoft’s AI robot that can describe images. The uses of this sort of capability are extremely broad, but marketing professionals especially must be foaming at the mouth. Other features such as motion detection and optical character recognition as well as superior speech-to-text capabilities will certainly have users flocking to the upgrades. Also exciting is the Hyperlapse developments, which Microsoft promises will enable time-lapsing of videos in Azure Media Services without frame limitations.
Chelsea Kerwin, July 26, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark Web meet up on July 26, 2016. Information is at this link: http://bit.ly/29tVKpx..
Sinequa: Now Seeking US Partners
July 25, 2016
Sinequa, a French search vendor, is hunting for partners in the US. The news appears in “Sinequa Partner Advantage Program Empowers the Channel to Capitalize on Leading Cognitive Search & Analytics Technology.” If you liked the title of this article, you will love the subtitle:
Company Launches New Partner Program to Drive Cross-Industry Adoption of Cognitive Search & Analytics and Address Growing Customer Demands
Keywords galore. What I noted was the euphony of “leading cognitive search and analytics technology.” A number of outfits are chasing the “cognitive search” pot of gold. Competitors include the champion in declining quarterly revenue IBM. Then there are the assorted machine learning folks at the Alphabet Google thing. Plus there are various and sundry deep learning initiatives appearing on a daily basis from the money crucible in Sillycon Valley; for example, Indico, MetaMind, Ripjar, Synapsify, and, my favorite, Idibon. I just love “idibon.” So many associations from ichibon to bon bon. Good, right?
Partners flock like Zika bearing mosquitoes when there is big money in a reseller/OEM/integrator tie up.
I learned from the Sinequa write up about Sinequa:
Sinequa continues to grow its partnerships with leading global systems integrators and value-add resellers (VARs) as well vendors of enterprise application, cloud and Big Data. In an effort to address rising customer demands from Global Fortune 2000 organizations for turning data into actionable insights, Sinequa extends its worldwide network with partners seeking to enrich their Big Data/analytics offerings in key strategic markets such as banking, defense and security, life sciences, manufacturing, utilities and government. The Sinequa Partner Advantage Program enables channel and service partners to quickly capitalize on the high growth opportunity in cognitive search and analytics. Designed to empower partners with certification programs, technical support and world-class training, Sinequa also offers partners performance-based incentives and marketing support programs…Certified partners access the recently introduced Sinequa ES Version 10. Powered by Machine Learning capabilities at its core, this ground breaking version helps deliver deep analytics of contents and user behavior, offering information with continually improving relevance to users in their work environments.
A point I think is important: Sinequa was founded in 2002. That makes the company 14 years young. Not quite a start up but agile enough when it comes to cognitive technology.
I assume that in today’s economic environment, potential partners will be swarming like the Zika bearing mosquitoes in the river marsh near my home in Harrod’s Creek, Kentucky. These critters seem to fancy my chubby, 72 year old body.
I have noted, however, that some vendors of search are having to work extra hard to close deals. Examples range from Big Blue in Union Square to SLI Systems in New Zealand and parts in between.
The idea of partnering is a good one. Endeca rose to its legitimate $100 million plus in search revenue with its carefully crafted partnering program. On the other hand, the Google Search Appliance partners continue to regroup because the wiser minds at Mother Google killed off the pricey Google Search Appliance. I treasure my print out of the GSA schedule with the five and six digit license fees for the wonderful GB 7007 and 9009 models. Imagine a locked down appliance for the price of a pre acquisition Autonomy IDOL license. Then when the document capacity of the search appliance was reached, a customer could license more Google Search Appliances. I found this business model interesting because taxi meter pricing is often an issue for chief financial officers who want to budget for certain products and services.
The upside of partnerships is that, as Endeca learned, unusual opportunities can be discovered. Once the deal is closed, the lucky partner has an opportunity to tailor the search system to meet the needs of the customer. Once up and running, life is good. Renewals, customization, consulting, maintenance fees, and other oddments make a search vendor’s life one of comfort and joy. The downsides include lawsuits, squabbles, and disruptions from competitors.
Worth watching how Sinequa maneuvers in the US market. Other French search vendors have found the costs and cultural issues a bit of a headache. Examples range from Antidot, Pertimm, and Exalead among others. Do you use Qwant?
Stephen E Arnold, July 25, 2016
Content Marketing about Bing Changing Lives
July 25, 2016
I love content marketing. Stories which contain a mixture of facts and other information are amusing. Consider “How the Power of Search Has Changed the Way We Live.” I use Bing. I also use Yandex, the Google thing, Unbubble, MillionShort (when it is online), Gibiru, and a number of other systems. No one search system duplicates the result sets of other systems. The write up blithely ignores this observation.
I learned that I could learn about search in a Microsoft white paper (yep, another content marketing thing) called “The Humanization of Search.” I assume Microsoft has abandoned its effort to co-opt the phrase “beyond search.” Nice try, folks.
You can download this write up from this link and watch a video. The write up is 18 pages of juicy fruit. I noted three statements:
- Voice queries are longer than text queries
- People ask questions when entering a search via voice
- Questions use who, what, how, when, and where structures.
Okay, take a moment to catch your breath.
Microsoft wants to be the big dog in voice search. I understand. The hitch in the git along is that the big dog seems to be cross town neighbor Amazon with its weird black speaker gizmo. Then there is the persistent problem of the Alphabet Google. Microsoft is in the game, but I don’t see the company pushing the Messis and Ronaldos of voice search to the second team for a while if ever.
Like IBM, the notion that saying things is much easier than delivering results. I find the parallel between IBM Watson cognitive computing marketing and IBM’s performance start evidence that talk does not generate sustainable revenues and rising profits. Microsoft may be dazzled by its white paper lingo, but the company has to demonstrate that its mismanagement of the mobile market is an exception, not the steady pulse of missing shots in front of the goal.
Read the white paper. Watch how the shift from search leads to marketing; for example:
As experiences across platforms become more prevalent, marketers need to familiarize themselves with emerging technology, as well as the massive growth opportunities that stem from search being more incorporated into everyday human life.
Confused. So was I.
Stephen E Arnold, July 25, 2016
-
- Subscribe to Beyond Search
Feature archive
News archive
-
