Semantic Search

From BC$ MobileTV Wiki
(Redirected from SmartSearch)
Jump to: navigation, search

Semantic Search (sometimes also referred to as SmartSearch) is any searching algorithm, querying technique or information retrieval system that is supplemented by Semantic Web technologies and methodologies such as metadata, profiling, similarity maps, clustering, RDF, OWL, Description Logics and Machine Learning algorithms. Semantic Search uses the science of meaning in language—instead of just searching keywords, it checks the context of the words to return more relevant results. Brooke Aker, CEO of Expert System USA, predicts that it will usher in the era of Web 3.0

10 Characteristics

A Semantic Search Engine should be able to handle:

  1. Morphological variations (tenses, plurals .vs. singular, etc)
  2. Synonyms with correct senses (cure = heal, fix, treat, remedy, relieve, medicate, etc)
  3. Generalization understanding (GERD, HIV, AIDS, ALS, MLS, are all types of diseases)
  4. Concept matching ("cure headache fast" should match to migraine medical condition & possible treatments, "fast" implies severity of condition and importance of alleviating severe pain quickly; also smart enough to warn that severe headache can be sign of more serious problems: stroke/aneurysm, or more basic self-curable problems i.e. dietary deficiency - high copper levels, calcium/iron off, etc)
  5. Knowledge matching ("swine flu prevention", swine flu = influenza strain = H1N1 --> location maps of known outbreaks, hygiene prevention tips, dietary & immune system boosting tips, etc)
  6. Natural language queries and "frame-of-thought" questions ("How do I build a tree house?" --> not a real house, playplace for kids, should be safe/durable... what types of wood/materials, tools & building supplies are best?, example blueprints/designs, step-by-step guides, what kind of trees does person querying have in their location to help prioritize/rank results?, etc)
  7. Pointing to uninterrupted paragraphs & most relevant sentence (not just link to a page but jump straight to the point, the "answer" or relevant section of the broader text; help minimize pagination-based ads)
  8. Entering queries freely, no special formats like quotes, or Boolean operators (these should be added automatically, not required to be entered by the "hacker/power-user")
  9. Operating without relying on statistics of their "overall user-base" (group-based generalizations/classifications, user behavior, or other artificial means)
  10. Detection & adjustment of its own performance (self-adjust based on how a particular user responds to the information, but don't "filter bubble" them, always show them or allow them to navigate to "the same thing anyone else would see for the same query")





Several players have emerged in the Semantic Search field, with no clear leaders, nor timelines for when their Alpha and Beta releases will gain mainstream attention and usability.




Up-and-coming semantic search company Hakia is launching a new social feature next week, called "Meet Others." It will give you the option, from a search results page, to jump to a page on the service where everyone who searches for the topic can communicate.



PowerLabs was the research division of Sillicon-valley start-up PowerSet which released a public beta for its Natural Language Search Engine in October 2007.

Powerset was acquired July 3rd, 2008 by Microsoft for $100 Million USD. Microsoft's intention is thought to be a desire to integrate the limited but sporatically impressive natural language processing capabilities of Powerset's Natural Language Search algorithms as a (much-needed) direct link between queries Microsoft receives through its MSN Live web search engine, and textual/display ads that it serves next to the queries.


TextWise creates the semantic Digital DNA of documents, revealing the fundamental meaning of text. Semantic Signatures® mine content to uncover the deeper meaning of text and creates a unique ’signature’ for every document it processes. Eliminate the need to find or categorize documents using ambiguous keywords or entities.

The TextWise Semantic Cloud provides a service for enterprises to retrieve real-time, highly relevant matches within their own private and secure dynamic content collections.


Text Used for Similarity Search section shows an example of the use of the filter API call<USER_TOKEN>/filter/web?uri=<URI>


Reputation Search.


Reputation Management.



Text extractions with graph for relevance and topic indication, plus tag cloud of most common words appearing in text.


Swotti is a new semantic search engine that aggregates opinions about products to help you make purchasing decisions. With Swotti, you can learn from the good and bad experiences of others as the site gathers together reviews and feedback from across the web and categorizes them to provide you with more information about the product you're interested in. What's unique about this search engine is that it uses semantics to do so.







<?xml version="1.0" encoding="utf-8" ?>
<spellrequest textalreadyclipped="0" ignoredups="0" ignoredigits="1" ignoreallcaps="1">

External Links


  1. What is Semantic Search? -- 10 Things that Make Search a Semantic Search:
  2. What is Semantic Search? (From ZDNet):
  3. PoolParty -- Semantic Graph DB - What is a Knowledge Graph?:
  4. Hakia launching new spin on social searching:
  5. Watson REST API:
  6. Building Semantic Web Based Applications with Watson:

See Also

Semantic Web | Search Engine | Freebase | DBpedia | Metadata | Tags