Our capstone project sees us utilizing machine learning to address an issue of “bad patents”, whereby new patent filings with a high level of similarity to older filings cause disputes which in turn cause significant wastage of time and money on litigation in court. If every new patent filing could be compared against an entire database of older patents before a decision is made, that would solve the issue, but the sheer number of patents in existence make this unfeasible if using humans. Machine learning, on the other hand, allows such actions to be fully automated, and thus presents a viable solution to the problem. Over the past nine months, we developed a multi-part software solution involving large-scale data retrieval and analysis, the implementation and training of a support vector machine learning model, and the creation of an interactive graphical user interface.
In this paper, I discuss in detail my specific contributions, which include creating a connected infrastructure and writing the graphical user interface using web tools. My work in those areas enhanced team collaboration, improved operating efficiency, and provided our target audience with an interactive portal for actually utilizing our work. Other critical technical work, such as the model training and data parsing, is analyzed inside the papers written by my teammates David Winer, Joong Hwa Lee, Dany Srage, and William Ho. I then follow up with an analysis of how our project incorporates key facets of engineering leadership, such as marketing strategy, competitive analysis, and ethical forethought. Our combined efforts resulted in a novel and robust software solution which we believe satisfies the needs of our target audience. In the long run, we hope our work will help steer the United States patent system back towards its original purpose of fostering innovation, while also contributing to increased public interest in machine learning as a tool to solve broad problems.
Title
Predicting Pad Patents
Published
2017-05-11
Full Collection Name
Electrical Engineering & Computer Sciences Technical Reports
Other Identifiers
EECS-2017-66
Type
Text
Extent
28 p
Archive
The Engineering Library
Usage Statement
Researchers may make free and open use of the UC Berkeley Library’s digitized public domain materials. However, some materials in our online collections may be protected by U.S. copyright law (Title 17, U.S.C.). Use or reproduction of materials protected by copyright beyond that allowed by fair use (Title 17, U.S.C. § 107) requires permission from the copyright owners. The use or reproduction of some materials may also be restricted by terms of University of California gift or purchase agreements, privacy and publicity rights, or trademark law. Responsibility for determining rights status and permissibility of any use or reproduction rests exclusively with the researcher. To learn more or make inquiries, please see our permissions policies (https://www.lib.berkeley.edu/about/permissions-policies).