Google proposes to apply AI to the generation and categorization of patent applications

The The Transform Technology Summits begin October 13 with Low-Code / No Code: Enabling Enterprise Agility. Register now!

Google says the patent industry should benefit from AI and machine learning models like BERT, a natural language processing algorithm that achieved industry-leading results when released in 2018. In a white paper released today ‘hui, the tech giant describes a methodology to train a BERT model on more than 100 million patent publications from the United States and other countries using open source tools, which can then be used to determine the novelty of patents and generate classifications to facilitate categorization.

The global body of patents is vast, with millions of new patents issued each year. It is also complex. Patent applications are on average around 10,000 words long and are meticulously written by inventors, lawyers and patent examiners. Patent filings are also written in language which may be unintelligible to the uninformed reader and highly dependent on the context; many terms are used to mean completely different things in different patents.

For all these reasons, Google believes that the field of patents is ripe for the application of algorithms like BERT. Patents, the company notes, represent enormous business value for a number of organizations, with companies spending tens of billions of dollars a year to develop patentable technology and negotiate rights to use the technology and patent offices. resulting patents.

“We hope that our [proposal] will assist the wider patent pool in its application of machine learning, including corporate patent departments looking to improve their internal models and tools with more advanced machine learning techniques, patent offices interested in leveraging cutting-edge machine learning approaches to aid patent examination and prior art research, machine learning, and natural language processing researchers and academics who may not have considered to use the patent corpus to test and develop new natural language processing algorithms, ”Google data scientists Rob Srebrovic and Jay Yonamine wrote in a blog post. “Patent researchers and academics who may not have considered applying the BERT algorithm or other transformer-based approaches to their study of patents and innovation.”

As VentureBeat recently reported, businesses aren’t the only ones benefiting from AI when it comes to patent processing. The United States Patent and Trademark Office (USPTO) built AI models for different categories of patents and then trained the models on the text of the patent abstracts. In addition, USPTO staff are using AI to process patent applications more efficiently. According to a spokesperson, the agency now uses a “leading RPA provider” to centralize its bot efforts and ensure an appropriate governance model and process that includes use cases, development, testing and security before the bots are deployed.

“We are working on adding AI tools to help get requests to reviewers faster and to help reviewers search for prior art,” said Andrei Iancu, US Under Secretary of Commerce. intellectual property, in an email response to VentureBeat in October. “We have also been active on the trademark side, exploring the use of AI to help find previous similar images and identify what we call fraudulent specimens. We are exploring the use of AI to improve the accuracy and integrity of the trademark registry.


VentureBeat’s mission is to be a digital public place for technical decision-makers to learn about transformative technology and conduct transactions. Our site provides essential information on data technologies and strategies to guide you in managing your organizations. We invite you to become a member of our community, to access:

  • up-to-date information on the topics that interest you
  • our newsletters
  • Closed thought leader content and discounted access to our popular events, such as Transform 2021: Learn more
  • networking features, and more

Become a member

Comments are closed.