This site uses modern web standards that aren't supported by your browser. For best results, please upgrade to Google Chrome, Microsoft Edge, or Mozilla Firefox.


How Semantic Search Eliminates Trial-and-Error Patent Research

|

international_patent_researchThe process of searching for prior art is extremely prone to errors and omissions. These could hinder the patent application later in the prosecution process or, worse, during future civil litigation brought by a party who discovers previously unidentified prior art—and hopes to use it to invalidate your patent.

Searching for elements of the name or descriptions of the earlier invention is, on its own, insufficient. The prior art you are looking for may not use any of the words or phrases you may expect to find. For example, prior art for a claimed invention of new door hinges might not use the words “door” or “hinge.” You could spend days hunting for inventions that comprise prior art for “a hinge-based anchoring system that facilitates the ready opening and closing of a frame-mounted door,” when the invention that is poised to derail your application is instead described as “a concentric set of metal cylinders allowing for rotational movement along a vertical axis.”

That leaves the unappealing option of a lengthy trial-and-error search through all the possible patent classifications that might encompass a prior art invention.

A matter of semantics

Thankfully, advances in search technology are making it easier to sift productively through the world’s mountain of patent data. Semantic search allows patent searchers to escape the linguistic guessing game when trying to identify prior art.

A semantic search does not confine the prior art search to exact or even partial matches on a particular set of search terms. Semantic search algorithms understand the nuances of language well enough to examine the phrase “a hinge-based anchoring system that facilitates the ready opening and closing of a frame-mounted door” and generate a cloud of comparable ideas.

Try the SenseBot search engine to get a feel for how semantic search works on Internet web pages in comparison to a traditional matching search engine.

Applied to a patent prior art search, a specialized semantic search algorithm could identify the word “hinge,” discern the context as a hinge for a door, and understand the definition of the word hinge well enough to predict the possible descriptors and combinations of words that could reasonably refer to either a hinge or a hinge-like device.

Being wholly unique

The burden is on the patent professional to make the case that the claimed invention is something that is wholly unique in the entire history of recorded human endeavor. Quite a standard to meet, it is only getting more difficult to prove patent uniqueness. More than twice as many patents exist today as there were just 30 years ago. In 2013 alone, the number of patents issued crossed the one million threshold for the first time, according to WIPO statistics.

Today’s advanced patent search tools can provide reassurance that you are not missing any critical prior art. The most powerful search tools combine deep search with a growing number of patent records for international patent authorities to ensure access to the most complete and up-to-date international patent research.



New Call-to-action