In the newest version of PatBase, a “Summarise” tool can be used to condense the Title/Abstract, Claims, or Description sections of a patent document to 1, 5, 10, 25, or 50% of its original length. If you don’t have a PatBase subscription, free options are also available. Back in January, 3 Geeks and a Law Blog highlighted 5 Resources for Summarizing Web or Other Content. Two of these applications, GreatSummary and FreeSummarizer, seem particularly promising for patent summation, since they allow users to cut and paste large blocks of text into a form and choose the number of sentences for the condensed output.
After the jump, we’ll experiment with the PatBase “Summarise” tool and the two free summarizing tools to see which application produces the best patent document summary!
PatBase’s “Summarise” Tool
First, I tested the new “Summarise” tool on PatBase. When a user is viewing the full-text sections of a patent document on PatBase (arranged under three sections: Title/Abstract, Claims, and Description), they can select the “Summarise” link to automatically condense the full text section to a selected percentage level (1, 5, 10, 25, or 50%). The user can them save or print this summary. Selecting the “View all” option will reveal the full text of the section to the user, with the summarized part of the text highlighted. I tried summarizing the claims section of US2004234820A to 5%, and the resulting summary appeared to highlight a single claim (out of 41 claims).
I then cut and paste the entire claims section of US2004234820A into GreatSummary.com. GreatSummary is a free online tool that allows the user to either paste a large block of text or the URL of a webpage into a form on the site, select a number of sentences for the final summary, and produce a summary of the pasted text or webpage. I chose to produce a five sentence summary of the entire claims section, and the resulting summary was rather long due to the length of a sentence that constitutes a single claim. The summary from GreatSummary.com seemed to be made up of five individual claims.
Finally, I pasted the claims section of the US patent application into FreeSummarizer.com, another free automatic summary tool where the user can paste a block of text and select the number of sentences for the final summary. I chose to produce a five sentence summary, which again resulted in a lengthy summary consisting of five individual claims.
The main problem with trying to summarize patent text within the free online summary tools is that the condensed output is based on numbers of sentences, and sentences can be rather lengthy within a patent document. The format of patent claims, especially, leads to complex and run-on sentence structures, so a summary based on a certain amount of sentences will result in a number of claims being incorporated into the final summary. It still may be easier to read five important claims than to read all 41 claims, but I personally preferred the summary output based on the PatBase “Summarise” tool. The PatBase tool focused on percentage of text rather than number of sentences, so the final output is guaranteed to be a specific length (in proportion to the original length). I also appreciated the option to view the summarized text highlighted within the full text, so the user can see where important keywords may be concentrated in the patent document.
The PatBase ”Summarise” tool produces the most concise summaries, since it condenses the text to a certain percentage instead of a number of sentences. Ultimately, though, the searcher should use these summary tools cautiously, since there is always the chance that the summary may overlook an important claim or keyword that makes the document relavent to the search. Neither the PatBase manual nor the websites of the free summary tools describe how exactly the applications identify which sentences or sections of text to extract, so the user may not be getting the most high-quality summary available. If a user wants a high-quality, hand-produced summary of a patent document, they should use the Derwent Abstracts from the Derwent World Patent Index.
Do you know of any online tools or database features that produce useful summaries of a patent document? Let us know in the comments!
This post was contributed by Joelle Mornini. The Intellogist blog is provided for free by Intellogist’s parent company Landon IP, a major provider of patent searches, trademark searches, technical translations, and information retrieval services.