Language Aware Search

Intelligent retrieval uses knowledge-based methods to capture and "understand" how you personally use language and syntax.

Although the vocabulary you use in your journal is large, it is a fraction of the number of words you can recognize when you read―which, in turn, is a fraction of the words in the dictionary. The terms you use when you update your journal are domain-specific―that is, specific to your worldview. Now & Then uses this domain to determine how you, personally, use language and associate one word (or phrase) with another. It uses terms and names to identify  the people, places, and concepts that are part of your worldview.

To understand your use of language, Now & Then combines language rules and language modeling including semantic parsing, term scaling, classification, and term weighting.

You initially train Now & Then to understand the basics of how you, individually, use words and syntax. As a new user, you "initialize" Now & Then's language understanding by ranking word attributes. You rank from 1 to 10 how "close" one word relates to another. This ranking is unique to every individual. Examples of term attribute ranking:

"On a scale from 1 to 10, with 1 being more closely related than 10, rank how closely you relate the term "project" to each of the following terms:
▪ work ▪ hobby ▪ community ▪ family▪ children ▪ enjoyment ▪ avoid ▪ list ▪ pressure ▪ play"

"Similarly, rank the term "region" to each of the following terms:
▪ U.S.A ▪ international ▪ neighborhood ▪ bay ▪ travel ▪ vacation ▪ fly ▪ sell ▪ local ▪ Mexico"

You also initiate your personal Synonym, Controlled Vocabulary and User-Thesaurus lists. The lists can be updated and modified at any time using Now & Then's feedback dialogue. These exercises provide Now & Then with a base level of understanding about how you use language―an understanding that it employs to retrieve contextual relevant information during searches and queries. This initial, or base level training id just the beginning―Automatically, with use, and through your feedback, Now & Then refines its understanding of your world and the words and phrases you use to describe it.

Your word attribute rankings are assigned ever-changing mathematical ordinates for their location in a multi-dimensional matrix. This matrix, or N-dimensional space, tracks associations and relationships and identifies clusters of concepts and patterns the among all your journal terms and phrases.      Language Aware search (continued)

see: The Intelligent Journal:  Context Aware Search   ▪   Query interpretation technologies ▪ Knowledge Based & Prompted search

home