Microsoft and Mikojo Trigger Semantic Winds across Search Landscape

January 28, 2010

Semantic technology is blowing across the search landscape again. The word “semantic” and its use in phrases like “semantic technology” has a certain trendiness. When I see the word, I think of smart software that understands information in the way a human does. I also think of computationally sluggish processes and the complexity of language, particularly in synthetic languages like English. Google has considerable investment in semantic technology, but the company wisely tucks it away within larger systems and avoiding the technical battles that rage among different semantic technology factions. You can see Google’s semantic operations tucked within the Ramanathan Guha inventions disclosed in February 2007. Pay attention to the discussion of the system and method for “context”.

image

Gale force winds from semantic technology advocates. Image source: http://www.smh.com.au/ffximage/2008/11/08/paloma_wideweb__470x289,0.jpg

Microsoft’s Semantic Puff

Other companies are pushing the semantic shock troops forward. I read yesterday in Network World’s “Microsoft Talks Up Semantic Search Ambitions.” The article reminded me that Fast Search & Transfer SA offered some semantic functionality which I summarized in the 2006 version of the original Enterprise Search Report (the one with real beef, not tofu inside). Microsoft also purchased Powerset, a company that used some of Xerox PARC’s technology and its own wizardry to “understand” queries and create a rich index. The Network World story reported:

With semantic technologies, which also are being to referred to as Web 3.0, computers have a greater understanding of relationships between different information, rather than just forwarding links based on keyword searches.  The end game for semantic search is “better, faster, cheaper, essentially,” said Prevost, who came over to Microsoft in the company’s 2008 acquisition of search engine vendor Powerset. Prevost is still general manager of Powerset.  Semantic capabilities get users more relevant information and help them accomplish tasks and make decisions, said Prevost.

The payoff is that software understands humans. Sounds good, but it does little to alter the startling dominance of Google in general Web search and the rocket like rise of social search systems like Facebook. In a social context humans tell “friends” about meaning or better yet offer an answer or a relevant link. No search required.

I reported about the complexities of configuring the enterprise search system that Microsoft offers for SharePoint in an earlier Web log post. The challenge is complexity and the time and money required to make a “smart” software system perform to an acceptable level in terms of throughput in content processing and for the user. Users often prefer to ask someone or just use what appears in the top of a search results list.

Mikojo’s Financing Gale

I also learned that this Internet search engine company secured additional lines of credit. The figure is in the $23 million range, which is healthy indeed. Mikojo, according to the news story on Yahoo:

Mikojo is a provider of Internet search services and technology. Mikojo is implementing technology that it believes will improve the search experience for consumers by providing mechanisms for users to specify search queries that integrate multiple data sources on the Internet. Based in Foster City, California, with offices also located in Los Angeles and Australia, Mikojo has a history of technology innovation in data integration and Internet search. The Company is incorporated in Delaware and trades on the OTC Bulletin Board under the symbol MKJI.

Google’s link to Mikojo references a “semantic search engine”, and the company’s splash page has the tag line “the intelligent search engine.” My hunch is that Mikojo is also indicating that it uses technology that understands content. Here’s what the company says on its About Us page:

Mikojo is focused on users who have a specific purpose or decision in mind when they utilize Internet search. For such users, Mikojo provides ways to search for criteria across multiple data sources by correlating and integrating these data sources together. Results are presented as tables, allowing users to understand the information more easily and render decisions rapidly. The company is dedicated to setting the standard for search technology and how people find information. As the Web becomes larger and more complex, finding relevant information efficiently has become increasingly critical to Internet users. Mikojo uses the virtual relation platform called TriggerWare as the foundation for its search technology. By innovating the proven TriggerWare© search technology and adapting it to the complex problem of Internet search, Mikojo helps users find what they need quickly and intuitively. The new search facilities are fully integrated into Mikojo’s standard search engine, thereby helping users refine their searches and quickly access the most pertinent and useful ways to gain the information that drives their insight and decision-making.

AltSearchEngines comes right out and says:

The TriggerWare™ query language, similar to SQL (Structured Query Language), provides a more semantically-based paradigm for search than a simple keyword search. As queries involve different tables from different web pages, answers to queries require semantic processing of information from different web pages. The company believes that no current search engine can integrate information from multiple web pages in a semantically significant manner. TriggerWare™ provides smart alerts over complex queries, enabling users to see the answer to their queries, and be notified when such information changes.

The Goose’s View

My hunch is that the semantic winds are picking up. Microsoft is mentioning semantics and investors have voted with their business savvy on the Mikojo’s method.

Several thoughts:

  1. Semantics is a term that needs to be defined and anchored in a specific technical context within a software system and method. The word alone has limited meaning for me. Others may be different in their understanding of “semantic”.
  2. Semantic systems can be computationally intensive. The explanation of semantic technology, therefore, requires some indication of content throughput on specific hardware, index latency, and scaling methodologies. “It scales, no problem” is not enough for me.
  3. Search is undergoing significant mutation. On one end of the spectrum are “one size fits all” systems. As I heard in Europe last week, many European companies prefer an SAP type approach with responsibility fixed on one company for enterprise software solutions. On the other end of the spectrum, azure chip consultants are in a tizzy over the use of search technology in specific problem solving use cases; for example, eDiscovery or customer support applications. The problem is that a two dimensional spectrum does not embrace the environmental changes that companies like Google, Microsoft, and Exalead are making.

I wonder if the semantic winds will pick up or die down or just come and go. Buzzwords come and go as do trends in search. Put on your ear muffs. Some winds can howl and chill one’s bones. Do not wear a jacket stuffed with goose feathers, please.

Stephen E Arnold, January 28, 2010

A freebie. I will report the inclement weather and the non compensation mode of this article to NOAA, an outfit that controls many things just not howling winds.

Comments

2 Responses to “Microsoft and Mikojo Trigger Semantic Winds across Search Landscape”

  1. Charlie on January 28th, 2010 7:27 am

    I’m afraid that we’re still far away from “smart software that understands information in the way a human does” – otherwise surely we’d all have robot butlers by now? 🙂

    Semantic search still depends on:
    – content being richly and correctly tagged (with the investment this requires)
    – decent NLP solutions – for each language – to extract semantic data from raw text
    – semantic nets such as WordNet – for each language

    However we do have a great deal of semantic search marketing…

  2. What is Web 2.0? on February 2nd, 2010 8:55 pm

    […] Microsoft and Mikojo Trigger Semantic Winds across Search Landscape (arnoldit.com) […]

  • Archives

  • Recent Posts

  • Meta