Ricci Dipshan , Legaltech News
The startup, which
was seeded by the National Science Foundation, uses semantic searches to find
hidden patterns between documents.
For everything outside of the deep web, lawyer
and layman alike can navigate today’s sea of information’s through search
engines, a critical piece of infrastructure for any web user or e-discovery
practitioner. But as data growth and complexity show no sign of slowing, are
today’s conventional search engines able to keep up?
One new entrant in the field is betting the time
is ripe for a new search approach. The recently launched Omnity knowledge
discovery and research engine seeks to revolutionize search engine capabilities
by focusing on more than just connecting a files and data to a few specific
terms.
“Traditional search engines are based upon
keyword search”, explained CEO Brian Sager to Legaltech News. “Omnity can
search using entire documents as queries, not just keywords. This avoids the
need to select a small number of words to represent a complex or nuanced
concept, as is often the case with legal documents.”
This semantics-based search goes beyond finding
related documents via links or citations, and instead finds what the company
describes as “hidden semantic patterns.” Sager noted that this process “uses
the rare text in a document to connect to the rare text in other documents.”
“The fundamental technology is based on
determining the shared, rare words in documents by first constructing for each
document a mathematical equation describing the statistical distribution of
rare words in that document, then matching documents by the similarities of
their mathematical equations,” he added.
Omnity markets the search engine to legal
professionals, noting that it was developed with advisory support from legal
experts, and pulls together “a wide range of documents together, including U.S.
issued and pending patents, US case law, SEC filings, courseware, news
articles, technical and scientific articles, which may help with patent
prosecution as well as technically-focused litigation,” Sager said.
Where the search engine can help, Sager added,
is in finding “the 99 percent of legal documents that are not directly linked
to one another, but are highly semantically related.”
Developed by research scientists and engineers,
Omnity’s initial seed funding came from the National Science Foundation, which
provided grants for its early development. Sagar noted that the search engine
has attracted additional private equity financing, but declined to reveal the
specific amount raised.
While Omnity is free in its public version for
all users, and in its academic version for those with an institutional email
address, it charges subscriptions fee for its professional and executive
versions. With these subscriptions, users can also add-on a legal bundle that
contains “all post-1976 issued and pending patents, U.S. Case law, and Federal
draft bills and legislation,” Sager said.
No comments:
Post a Comment