How Valuable Is Your Idea? The Mechanics For Valuing Intellectual Property
Article written by Loci developer, Mike Aebig.
You’re probably wondering how inventions are analyzed by Loci. What makes it possible to sift through over 10 million patents every single time for each analysis that’s conducted? Loci’s latest release, Invention Analysis, helps new inventors determine the uniqueness and demand for their ideas before they spend money trying to secure legal counsel or file a patent. In fact, with the advent of our latest technology and analytics, inventors may decide to forego the traditional route of patenting altogether and sell their invention through Loci’s blockchain-powered ecosystem. It’s that useful. We’re pleased to invite you to read on to see what’s behind the curtains of Invention Analysis below.
The intensive process begins when Loci’s Invention Analysis first receives the user’s idea (also known as an invention), submitted in a plain text format. Keywords are extracted from the plain text and then searched against the USPTO and EPO online resources to gauge which Cooperative Patent Classifications (CPCs) the user’s idea will fit into. A CPC stands for a Cooperative Patent Classification- a system that both the USPTO and EPO use to identify where a patent exists in the entire landscape of inventions. To an inventor interested in bringing life to an idea, understanding various classifications and how they may apply to their idea is an incredible advantage.
This also allows an inventor to consider alternate use cases for their idea that they had not previously considered. Enter the Loci lightbulb moment. The most powerful part about this feature of Invention Analysis is that at this point, not only does the inventor see anticipated areas of classification where his/her invention falls, but they also are presented with a completely new possible set of classifications and use cases for their idea. This feature is quite literally the ‘aha’ moment that so many inventors fail to achieve while conducting their own searches and offers a completely new and incredibly valuable perspective on how an idea or invention can be applied in the real world.
The plain text is again analyzed to create combinations of terms, CPCs, and other data, to create a large Conjunctive Normal Form expression. These expressions serve to compare multiple sets of clauses or search terms used as a database query parameter and sent over to the data source.
An invention is typically several paragraphs long, with approximately 40 keywords and a yield of 7000–27000 unique patents. The patents are then ranked based on relevancy with respect to the submitted invention through the implementation of a Vector Space Model. A Vector Space Model is a simple model that represents documents and queries as vectors and allows partial matching, ranking documents according to their possible relevance, and computation of a continuous percentage of similarities between documents. Upon ranking the patents from the database, the concluding score is spread amongst their CPC membership, and a per CPC sum is computed. This yields a CPC relevancy score, and a set number of the most relevant CPCs are returned to the user. Finally, the most commonly occurring keywords in the database response that are not present in the original invention text are returned to the users in a Keyword Cloud, an aesthetically pleasing front end GUI that presents “missing key words.”
The overall results of this process help users identify where in the IP landscape the invention or idea fits, determine how the invention is unique, discover industry terms for the patent space within which the user is working, and view other types of applications for the same technology.
This is only the beginning of creating a modest size runtime dataset and identifying features for building models which we will train for making predictions which has not been possible before.
The technology we are building today has the potential to change the way the world invents and ultimately the way we will all value ideas. It is about bringing access to masses that previously didn’t have the resources to innovate. We are excited to share this journey with you and look forward to unveiling more updates as we forge ahead.
Source Citation:
G. Salton, A. Wong , C. S. Yang, A vector space model for automatic indexing, Communications of the ACM, v.18 n.11, p.613–620, Nov. 1975
