Michael Plishka, a self-described “innovator” with 15 patents issued and pending — mainly in the field of medical devices — writes in his blog “Zenstorming — Where Science Meets Muse” about a web site (http://braindot2.com/) that offers intriguing artificial intelligence and pattern recognition web tools. Specifically, he cites the algorithmic patent search tool developed by Lipa Roitman, a chemist and founder of The Third Ingredient Labs, a product development and consulting lab with offices in Dayton, OH, and Netanya, Israel. The web tool not only shows patents in a particular field of interest but also allows users to calculate an approximate value for each patent.

For illustration purposes, Plishka ran a search for U.S. patents using the terms “camera” and “flexible” in the abstract. The results are displayed in the following graph:

Click to enlarge

If a box is outlined, the patent in that row belongs to the subclass of the column, Plishka explains. The number in the intersection of each column and row equals the number of subclasses in the patent plus the number of patents from the search in that subclass, he adds. The higher this number, the more valuable the patent. The shade of blue used in each box helps to carve out islands of similar groups of numbers. After running a search, users can check on patents they’d like to examine in more detail and click a button to conduct a sorting operation. “I even found some patents that didn’t show up when doing other web-based searches,” Plishka reports. The tool also depicts the relationships between various patents and classes and shows “holes in the patent landscape,” Plishka says.

Source: ZenStorming

Posted September 9th, 2009 under Tech Transfer

Write a comment