Technology

The Synthetix Difference

Why is current search and navigation technology insufficient in meeting the needs of the legal community in dealing with e-discovery?

First, traditional search technology is outdated. Traditional keyword search identifies all instances of the word in the document. This technique is useful for finding specific information in a small database, but in large databases, and particularly on the Internet, a simple keyword search is ineffective because almost any search will result in an unmanageable number of hits.

To work around this problem, Internet search engines such as Google capture and rank the results of searches its users perform in a sort of popularity ranking. But for this to work, this search approach needs legions of users, making it less useful for searches within an enterprise, or searches within a closed, private collection of records.

Traditional search techniques fail when it comes to e-discovery applications, because of the forced tradeoff between recall and precision. Recall is how much information is returned from a query - everything that matches, no matter how relevant, and precision is how closely those search results match the query. High recall can overwhelm the user with more records than could possibly be reviewed, any one of which may or may not contain relevant or critical information. High precision reduces the number of records but at the risk of excluding results that could be highly relevant. Having one without the other compromises search results since, in either case, important information is overlooked or simply never found. This limitation can translate to a significant cost or liability in a mission-critical application such as litigation, compliance or business intelligence.

Other types of search, such as concept search, require specific taxonomies which need to be updated on a regular basis, and which are often subject to someone else's interpretation of what a good result might be. In an attempt to improve on this, some concept search technology 'learns' from users to try to increase accuracy on future searches. This can mean that a different result is returned on the same query - which can be highly problematic in court.

The Syngence Solution - Synthetix Linguistic Pattern Matching Technology

Synthetix technology is different, and for use in e-discovery applications, measurably better because it frees users from having to choose between high recall and high precision by providing both -- with ranked relevancy.

Syngence is a pioneer in the development of linguistic pattern-matching technology, which treats the individual words in a document collection - and the unique combinations of words that comprise the essence of each particular document - as mathematical values.

Syngence's patented technology examines every page in a collection and creates a digital fingerprint of sorts.

Why is this important? Because it lets users - even those with limited computer skills - search using longer, natural-language search queries. Users can copy and paste - or simply create - a sentence or paragraph and Synthetix will deliver comprehensive results quickly. In addition, Synthetix is less sensitive to errors, OCR noise and misspellings, returning highly relevant results quickly. With Synthetix, only relevant documents are returned and no relevant document is overlooked.

Our Synthetix linguistic pattern matching technology is designed to allow users to