The machines are taking over

Are you ready for R2D2 and C3PO to take over? No, Yes, Maybe or Does not compute?

Ready or not, Machine Learning and Artificial Intelligence (AI) are steadily taking over every aspect of routine life including automation in the field of IP. In all aspects of life, the AI revolution is here to help us, even though the solution is removing the human broker from the workflow.

Over the last 20 years, Machine Learning and AI have helped to change and shape the IP patent information industry. OCR (Electronic Optical Character Recognition), Machine Translations and Semantic Search are three areas where technology has already helped.

The current status

OCR processing of patent PDFs and printed patent documents has been with us for a few decades and is a mainstream technology.

• Machine Learning – Mature Technology

• Benefits – Accuracy and Speed

Machine Assisted Translations are again a mainstream technology making global patent documents available in your choice of language format including on-the fly translations.

• Machine Learning – Mature Technology

• Benefits – Accuracy and Speed

OCR and Machine Assisted translations are standard background technologies. Deployment of these technologies was important due to the increasing volumes of global patents ensuring quicker access to the latest publications for patent professionals and organizations.

Semantic Search and the application of this technology for IP opened up patent searching to non-patent professionals. Semantic Search has made it possible to search patent data effectively without having to think for many hours how to formulate a Boolean search query or what the IPC or CPC code is for a washing machine.

• Machine Learning – Maturing Technology for patent searching

• Benefits – Precision (Accuracy), Relevancy and Speed

Semantic Search for patent searching is a developing technology. It will continue to improve and, just like the giant on-line search engines, the algorithms will “learn” with human feedback to improve the patent search results for users.

The focus now for both patent offices and organizations is managing the huge volumes of information with an increasing focus on technology and data scientists to help manage IP workflows.

How patent offices are evolving

Major patent offices are implementing technology including machine learning and AI as the future for their patent, trademark and design application workflow solutions. The Japan Patent Office have announced publicly that it is investing in the use of artificial intelligence technology to automate processes such as screening patent, trademark and design applications.

This is great news for inventors and those seeking IP rights. Getting granted status faster or having trademark or design rights is what the game is all about. The Japanese economy too will benefit from securing rights for their inventors and Japanese based companies more quickly.

Patent offices also provide external links to the IP office data to allow external access to the public data. Open access to patent office data is also going to be an important ongoing trend. The United States Patent and Trademark Office ”Open data and Mobility” (API) project has made available, via an API access, the public data.

The European Patent Office also has the “Open Patent Services (OPS)” API.

Organizations are also changing their approach to information management. In the past, patent data has been generally standalone in standard solutions. This approach in the 21st century is flawed and many organizations require solutions supporting data integration and/or data fusion. The latter solutions allow the integration of both structured and unstructured data to intelligently build relationships between multiple data sets. Organizations are focusing their IT developers and data scientist efforts to understand the possibilities of:

• Cloud computing,

• Integration and data fusion of IP patent data with internal and external datasets,

• Machine Learning and AI, and

• Visualization of very large complex datasets with graphical navigation.

Cloud computing offers a quantum leap in capability with access to a “state of the art” technology and database environments combined with huge processing power. Google, Amazon, Microsoft and Oracle all have big data solutions with advanced technologies including Natural Language Processing, Text Mining, Image Processing and Recognition, Analytics and more available in their respective Cloud computing environments. The price though is a trade-off between the potential performance benefits and having the data outside of your own environment.

Organizations not willing to host data in the cloud just yet do have an ever-growing list of locally deployable technology options. These include faster database engines and integration solution technologies. Some of the technology of the Cloud is also becoming more openly available through standard platforms with increasing numbers of vendors offering text mining, data fusion integration, analytic and visualization solutions including local deployments.

One real life vision of the present and future possibilities is a data fusion solution being built by the EU Commission project known as the “Tools for Innovation and Monitoring.” See the beta public platform here.

The future

So what is on the IP horizon as we push the boundaries of possibility and what technology can achieve? The biggest potential impact on patent workflows will be in the evolution of:

• Deep learning neural networks, integrations with data fusion and analytics platforms

• Image recognition processing and searching for:

  • Patent drawings, images, drugs, biomolecules, chemical and polymer structures
  • Trademarks and Designs (FYI - already happening)

This is all before we even contemplate an Intelligent Personal Assistant as the software broker for patent searching or specialist services.

The future of the IP business is certainly technology innovation based. The challenge for IP vendors is to marry both high quality content and innovative technology in 21st century solutions aligned to increasingly sophisticated user requirements.

One thing to keep in mind is that no matter how good any solution looks or how cleverly designed, the quality of the underlying data, including the patent data, must be at the top of the requirements list.

Now where is the number for the IT department…?