Asknet.ru: A Semantic Search Engine
January 27, 2014
We learned about a new semantic search engine. A public demonstration is available at http://www.asknet.ru/EN/index.htm. Some chatter about the system appeared on LinkedIn. Like many of the next-generation search systems, there were some questions and comments from the “experts” who participate in the LinkedIn search discussions.
According to the Web site:
AskNet search technology is its main product and the focus of the commercial licensing and development. The search engine combines the speed of an index with the functionality of linguistic analysis. The AskNet search engine reverses the search result process. Traditionally, search engines provide links to large numbers of documents that contain reviews. They leave the users to hunt their answers in thousands of pages and millions of words. AskNet`s linguistic analysis makes it possible to provide meaningful answers to searches as quickly as traditional search engines. No linking required!
One LinkedIn expert pointed out:
AskNet Search ( online service asknet.ru) is the demo version. Not all algorithms are implemented for asknet.ru. All of them are implemented in the enterprise search engine. AskNet Search realized metasearch functions using snippets from Google. These text snippets are not whole sentences. Therefore, the quality of linguistic search AskNet Search could be better, when it used sentences for search, rather than a snippets from Google.
We suggest running some test queries and determining if the system delivers useful results. Keep in mind that a technology demonstration is usually set up to make it easy to get a “feel” for the basics of a system.
With regard to semantics and analytics, the supporters of today’s hottest technologies often are like supporters of the Liverpool football team. The coach is usually wrong and one or two players are terrible. The team concept, however, is one to support to the death. Rational? Nah. Part of today’s standard operating procedure? You bet.
My view is:
- Many vendors are recycling old algorithms with a Project Runway touch up. The basic design, however, is recognizable as a cute little red carpet number. Innovation is a bow or a tuck.
- Some so-called experts (folks I describe as poobahs, azure chip consultants, of people with a dog in the fight) see their clients’ products as truly wonderful innovations. The notion that a researcher in 1980 hit upon a method and created a product based on that method is of little or no consequence. Who cares what Julius Caesar said after the battle of Alesia. Ancient history.
- Prospects may not be looking for a better search solution. Prospects may be looking for a system that is less of a problem than the incumbent solution. Therefore, the procurement team is trying to keep their paycheck, not revolutionize information retrieval.
- Many systems work only if the user knows what he or she is looking for. Predictive search (go with the search history and the norm for a cluster) is good enough. Who has time to do deep dive research in today’s rush-rush-rush business climate.
The buzzword blizzard makes it difficult to figure out what system delivers what. I know I am easily confused, and my hunch is that others may face the same hurdle. Will Sochi feature a confusion jump involving leaps of faith over search vendor claims?
Stephen E Arnold, January 27, 2014
Chocolate Toothpaste and Search
January 25, 2014
The ArnoldIT Overflight snagged this headline on January 16, 2014: ‘Chocolate Toothpaste’ Prevents Tooth Decay. Unusual news can be entertaining. Then on January 24, 2014, I spotted this headline: P&G’s Chocolate Toothpaste: Innovation or Desperation? The source of this story was not a secondary information source. The Innovation or Desperation item appeared in Bloomberg Businessweek.
Here’s the quote to note:
The line, which P&G (PG) promises to start selling soon, comprises three flavors: “Mint Chocolate Trek,” “Lime Spearmint Zest,” and “Vanilla Mint Spark.” Here’s how the Crest marketing team describes the new paste: “It’s a whole new world of deliciousness for toothbrushes everywhere.”
The hook for me was not “chocolate toothpaste prevents tooth decay.” This assertion reminds me of marketers who assert that a particular search system delivers value or understands human discourse. The problem is that the association of “chocolate AND tooth decay” is easier for me to grasp than “chocolate PREVENTS tooth decay.”
With search, value is difficult to connect to search. Search costs money, generates more work because documents have to be opened and read, or creates a willingness among busy users to assume that a search result is a correct result.
The Businessweek story connects “innovation” and “desperation.” Marketers have hit upon a product innovation that will make some influencers go for the chocolate toothpaste.
Search and content processing vendors have been following this path for many years. Not only is the uptake of new jargon standard operating procedure for search vendors, the consultants and experts working in search are turbochargers of constant exploration of ways to make search have sizzle. Search is analytics, taxonomy, knowledge, Big Data, etc.
I recall that in one of my university’s required classes, one professor insisted that ancient people cleaned their teeth with twigs. I also know that search methods from the years before “smart software” worked as well.
Progress in search and retrieval, like the chocolate toothpaste innovation, are not “innovation and desperation.” The juxtaposition of attributes is another indicator that the disconnect between expectations and reality is a characteristic of business today.
Will chocolate toothpaste work better than “regular” toothpaste? Will the new search system work better than “regular” search systems? Under specific test conditions, it is possible to “prove” efficacy. But in the real world, toothpaste like search has a baseline of performance. Wordsmithing, odd juxtapositions, and cleverness cannot be confused with Daliesque novelty.
Stephen E Arnold, January 25, 2014
Nuance Offers New Vocal Search
January 23, 2014
People have been waiting for years when they could go into their homes or work place and simply say a vocal command, then a computer would respond with the correct information. Nuance Communications is taking the world a step closer to that reality, says 4 Traders in the article: “Nuance To Start A Whole New Conversation With Launch Of New Dragon Assistant For Intel RealSense Technology.” Nuance’s new Dragon Assistant is available for computers that use Intel Real Sense technology and even better is that tablets can now use it.
Dragon Assistants comes with two personalities, British Butler and American female. Users can have an ongoing dialogue with their assistant and can ask it to do a variety of activities using vocal commands. Users will also be available to ask questions and receive answers. Dragon Assistant will have a knowledge pool supplied from over 170 content providers.
The article states:
” ‘Together with Intel, we are changing the way people experience their devices, starting a whole new conversation,’ said Michael Thompson, executive vice president and general manager, Nuance Mobile. ‘This latest version of Dragon Assistant is transformative. People can have an interactive dialogue that is natural and intuitive, with a voice assistant that listens and understands the context of conversation – just as you would expect from a personal assistant powered by Nuance and Intel.’ “
Expect there to be kinks in the Dragon Assistant. Developers have been trying to create an intelligent and human-like assistant for years, each project brings us one step closer. I presume, however, we still have not brought it up to Star Trek standards yet.
Whitney Grace, January 23, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
Elasticsearchs Facebook Happiness and Search Competition in 2014
January 21, 2014
A holiday leftover we’ve found at the Elasticsearch Blog has us contemplating the open-source search race for 2014. The site shares video from a recent event in which a Facebook representative discusses his company’s use of their product in, “Facebook & Elasticsearch: for Your Holiday Viewing Pleasure.” Is Elasticsearch surpassing Silicon Valley-based LucidWorks?
The post introduces the video:
“So, without further ado, we bring you this video from the inaugural Elasticsearch Silicon Valley meetup, in which you’ll learn more about Facebook’s use of Elasticsearch, including:
- Facebook’s migration from Apache Solr to Elasticsearch
- The company’s use of Elasticsearch to power internal search for developer tool sets and libraries
- How Elasticsearch powers Facebook’s Community Help Site
- And much, much more on their use case.”
The video is over an hour long, and full of good technical information, if that’s your thing. But the first two minutes summarize why Facebook prefers Elasticsearch over the competition. (The company had previously tried using Google Enterprise Search and Apache Solr and found each lacking.) Below the video, the post links to a webinar on getting started with their product. Formed in 2012, Elasticsearch is based in Amsterdam with offices in the U.S., the U.K., France, Germany, and Switzerland. They are also hiring as of this writing; that’s usually a good sign.
We heard that OpenSearchServer, another open-source search vendor, snagged the Le Monde account from Sinequa. If true, there seems to be competition between open-source search vendors and non-open-source search systems as well as among open-source search vendors.
Contention and competition. The year 2014 will be fascinating.
Cynthia Murrell, January 21, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
Search Application Perspectives For 2014
January 20, 2014
With 2014 well under way, search experts are trying to predict what will happen for enterprise search. Search Appliance World has an article that takes a look on enterprise search in the past and future called, “The New Search Appliance Landscape: Reflections And Predictions With MaxxCAT.” Basic search commands that come in out-of-the-box system are old school and do not provide the robust solution enterprise systems need.
Search appliances became enterprise users’ favorite toys and everyone had to have the Google Mini Search Appliance, but those days are gone. Other search developers, such as MaxxCat, stepped up to the plate.
The article states:
“ ‘In 2013, we saw a lot of the fallout from that as customers realized they couldn’t replace their Google Mini appliance and went looking for viable alternatives that weren’t $30K. For us, this lead to a huge boost in sales of our entry level appliances and even some additional sales of our enterprise series appliances,’ MaxxCAT Director of Marketing & Sales Chris Whissen told Search Appliance World.”
The MaxxCat developers were interested in exploring new markets their search appliance could expand into. The company is also big on customer service and ensuring that clients know they are valued. The biggest endeavor being made, though, is offering MaxxCat’s clients an efficient solution to solve their search problems and to encourage more competition in the search application market. Google is no longer the small player, but some of its solutions have grown too expensive for its former clients. New companies like MaxxCat keep the market fresh and offer up new ideas.
Whitney Grace, January 20, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
A User Friendly Search Tip from MarkLogic
January 20, 2014
Open source developers rely on code libraries their fellow developers have already written to complete their projects. This makes search engines and open source communities an invaluable tool. Blogs are another place to locate helpful code and the Developer Notes details one developer’s notes on XML, MARKLOGIC, XQUERY, XFORMS, XSLT, SCHEMA, JAVA, HTML, CSS, JAVASCRIPT, XSSI AND OTHERS.
By searching through the blog’s archive, we found this old post entitled: “MarkLogic: Techniques For Querying In-Memory Fragments Using CTS:Contains.”
Here is what you will find in the post:
“This snippet demonstrates the use of cts:contains and cts:element-attribute-word-query on an in-memory fragment (something that has been stored in the Expanded Tree Cache using a let statement).”
As any kind developer who pulls from open source, the author posts the code for anyone to use in their project. We occasionally find neat little tricks like this tucked away in the Internet. Sadly, many of them can get lost and are left in the hidden Web, which is we rely on deep Web crawler content wrangler. Developers need a robust search engine to find good code. Sometimes the big guns like Google do not work.
Whitney Grace, January 20, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
Basic Search Tips for SharePoint
January 16, 2014
The article titled Not Getting The Search Results You’re Looking For In SharePoint? on Microsoft Office divulges tips to help users who are having trouble with SharePoint, the content management platform from Microsoft. The basic issues it mentions relate to too many or too few results. For those receiving too few results, tips include generalizing search terms and ensuring your settings aren’t blocking the results. For too many results, the article suggests using Advanced Search.
This is what the article suggests for no results regarding syntax use:
“If an error message tells you to make sure you’re using the proper syntax, the search system interprets your search as a KQL query, but finds that there’s something wrong with the syntax. Check that you’re using the right syntax, and particularly the right number of parentheses and double quotation marks. If you want to search for a phrase that contains a parenthesis or double quotation mark, make sure that you enclose the entire phrase… in double quotation marks.”
You can also consult the Keyword Query Language (KQL) Syntax Reference for querying in KQL. However, we have a hunch that this advice is not too useful if the content is not in the index or some other system level issue is a problem. Consider the article more of a basic troubleshoot guide, not a comprehensive directory.
Chelsea Kerwin, January 16, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
Google Search Appliance VS SearchBlox Price and Indexing Limit Comparison
January 16, 2014
The article on SearchBlox titled Google Search Appliance Price Comparison with SearchBlox explains the slippery pricing data available to consumers on Google Search appliances. The article states that for a more limited document storage space Google charges $30,000 while SearchBlox, for unlimited storage, charges only $5,000 (but these numbers are only approximations). SearchBlox also offers more constant support and maintenance than Google, making it a very appealing option in the world of intranet or Web site search.
The article explains:
“SearchBlox provides the option of seamlessly moving away from the Google Search Appliance without skipping a beat. In addition to the cost-savings and feature comparison, scalability of the solution is something to consider given the explosion of content. SearchBlox scales both vertically (by adding more CPU/RAM to the existing setup) and horizontally (by adding more search servers that can be run in a cluster) without disrupting your architecture.”
SearchBlox even allows for Google administrators using XLS with a “faceted search plugin” that promises not to disturb the infrastructure. Allowing users to index unlimited documents certainly beats Google’s 500K indexing limit. A quick check of the GSA Advantage site shows that the Google Search Appliance is a significantly more expensive alternative to the open source based SearchBlox solution.
Chelsea Kerwin, January 16, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
A Sales Pitch for HP IDOL
January 12, 2014
Conceptual search allows users to search by concepts and ideas within information rather than basic keywords and phrases. Great idea, except that that the idea of conceptual search has been around since 1999. HP is touting it as a entirely brand new idea in the article, “Analytics For Human Information: Optimize Information Categorization With HP IDOL” posted on its own Web site. Rather than break directly into the “new” conceptual search, we are given the even better glittery term “categorization.” HP IDOL, using ExploreCloud-an SaaS solution for analytics and sights, offers an auto-categorization feature marked as a time saver and productive tool.
HP describes it as a magic tool:
“Powered by HP IDOL, ExploreCloud helps you uncover insights across all channels: web, mobile, social media, email, contact center, database, and storefront, so that you can organize and quantify content in a consistent, objective manner, resulting in data that is more accessible and consistent. And you can maintain existing legacy taxonomies and/or enrich them with contextual understanding. When you go beyond the limitations of what keywords can help you do, your whole world opens up. You can also discover the “unknown unknowns,” or topics you did not know to look for in the first place.”
The article stresses that regular keyword searching is far from abandoned, but its limitations are stressed. Keyword search’s weaknesses are addressed to the point of stating the obvious, and then it turns into a sales pitch for HP IDOL. Little is said about what exactly HP IDOL can do, other than organize data. HP, please tell us something we do not know.
Whitney Grace, January 12, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
More Watson Nuggets
January 10, 2014
Fast Company published “IBM’s Watson For Business: The $1 Billion Siri Slayer.” The write offers some nuggets of information that convert Watson from search system into the next Apple or Google. Frankly I find this notion somewhat amusing.
The story reports this interesting assertion, “IBM wants to transform Watson into a Siri for business.” Quite an analogy.
I also noted these items:
- Stephen Gold is the vice president of IBM Watson Solutions
- Watson Discovery Advisor will be a product/service for publishing, education, and health care
- Watson Analytics Advisor appears to be an interactive analytics solution
- An ecosystem will be built around the Watson Application Programming Interface and “the Watson headquarters will also include space for a tech incubator for startups building Watson-based apps”
- Watson will be deployed on Softlayer, an IBM cloud computing service. Apparently some eager Watson prospects have an appetite for Softlayer’s delivering Watson.
I marked a quote to note from Mr. Gold and Fast Company:
Watson for Business is “one of the top innovations in IBM’s history” and it could even be the biggest IBM innovation since the IBM PC.
IBM seems to have made a different executive available to the Wall Street Journal and the New York Times. My hunch is that the cheerleading will continue for a while.
Meanwhile where’s the online demonstration of Watson’s functionality? I want to see how the system compares to Hewlett Packard’s Autonomy technology, check out the visualizations to see if they are different from IBM i2’s, and figure out if the analytics are recycled SPSS functions or something different.
Stephen E Arnold, January 10, 2014
-
- Subscribe to Beyond Search
Feature archive
News archive
-
