Top 8 Trends from PIUG 2017

Patents Tags: Intellectual Property, Patent Conference, PIUG Annual Conference

By: Devin Salmon, Patent Analyst,

On May 20 – 25, the 2017 Patent Information Users Group (PIUG) Annual Conference was held outside Atlanta, GA. This conference is a place for patent researchers and other patent professionals to share knowledge and gain new skills.  It is also where the software/solution providers come to show off what’s new in their tools, which is an opportunity for attendees to learn about all the different options available.  This year the focus was on “The Complete 21st Century Patent Searcher-Addressing Our Skills Gaps.”

Here are eight trends I saw emerging from the presentations and from speaking to others at the conference:

  • Deep learning and neural networking is the next big thing. In the patent space, it is being used in many ways including improved translations, classification, and search.
  • Not all semantic engines are the same. Ask yourself: are they statistical models, artificial intelligence, neural networks, document signatures, deep learning, or a combination? How are they trained/tuned to assist with retrieval of patent data?
  • The big question is how do we as humans fit into patent analysis with the newer technologies available? Currently, the best practice appears to be a mix of machine learning and guidance from a human.
  • Not all translations are equal. There are old versions that were direct word-for-word translations, newer sentence based, and neural networks.  Questions around how translation affects the terminology used to describe new inventions still remains.
  • Clean data is king: every tool ends up having essentially the same data from all the patent offices. How they extract it, clean it up, and what additional features are added are going to be the keys to searching and using newer analytics tools.
  • Determining the “correct” assignee is still a major problem. Furthermore, how do we define “correct.”
  • Typical visualizations can be broken into two groups: those used for explanation and those used for exploration. The future lies in the creation of visualization tools that can be used to make decisions.
  • Tools for classification, subject grouping, and tagging have room for improvement and future versions should likely capitalize on the use of machines either to augment or replace a human.

Overall the view point from the conference attendees seemed to be that machine assistance is here to stay for the long term and is creeping into many aspects of patent search and analysis. We as patent information users need to keep up to date with the new technologies and embrace the doors they may open. We need to be willing to try them with an open mind and take advantage of them, rather than be left behind.