Google stoking the patent crowdsourcing fire

The patent "crowdsourcing" topic, although addressed before, appears to be gaining some momentum.  We know that Peer-to-Patent offers access to its website to spur the patent crowdsourcing initiative.  As suggested by Peer-to-Patent, Eric Schmidt, and others, crowdsourcing the patent review process may help the PTO avoid issuing broad patent claims. One problem still remains - how to apply crowdsourcing to the patent landscape whether at the PTO, for litigation, or in reexamination.

For example, Article One Partners is in the process of applying crowdsourcing to prior art search.  Peer-to-Patent applies crowdsourcing to prior art selection, with small doses directed towards analysis.  Prior art analysis, or tools to aid in analyzing the prior art specifically against claims, remain in the background of these efforts.  For crowdsourcing the actual prior art analysis, there's not much out there.

Although Lithosphere Software offers a set of patent tools directed to prior art analysis linked to patent claims, these tools are primarily designed for patent litigation to create invalidity contentions and invalidity charts.  Schmidt's suggestion below is a good one, but very labor intensive, and still is not directed at how to organize the analysis.

Schmidt suggested that the real problem with the current patent landscape in the U.S. has longstanding causes; in the early ’90s and early 2000s, he said, there were a lot of patents issued that “were very broad”, and that patent clerks are now spending a lot of time combing through and invalidating these older patents. The best way to address this problem, he said, is to take the patents as they’re published and crowdsource them. The best way to potentially curb these patent wars, then, would be to publish the patents publicly and allow “everyone to comment to see if there’s any prior art”, he said.

via Google Chairman Eric Schmidt Weighs In On Patent Issues: They’re ‘Terrible’ | TechCrunch.

We have considered how to extend our patent tools, possibly simplify them, to make them more appropriate to non-patent lawyers.  The gap in crowdsourcing patent and prior art analysis right now is how to allow "everyone to comment to see if there's any prior art."  Once prior art is identified, the next logical step is matching the prior art to the claim language.

Once you check out Lithosphere Software's platform, let us know how you think we could simplify our platform to improve it for use by the general public to achieve that elusive "how."

Crowdsourcing your invalidity contentions and prior art searches

In this post we'll consider a topic I've thought about for some time - crowdsourcing.  The concept is not new and has been applied to photography, social networks, software coding, and in a very limited way, prior art searching. Back in 2002, we were confronted with the problem of organizing a large set of prior art references.  Having been involved in a database project that organized IP assets, the database concept seemed appealing.  We searched for a commercial solution to our problem, but came up empty.  We thought about what we really wanted to accomplish and settled on a goal of extracting and retaining the most relevant information from each prior art reference so that the information could be quickly searched.  We designed a tool for that purpose in 2002.  The tool performed the retention function very well.  Because of the way we configured the database, extracting relevant information via search, beyond a list of references, was difficult.  In 2003, we developed a Microsoft Word scripting tool that crudely integrated with the database to generate a form of invalidity charts.

So what does this have to do with crowdsourcing?

We've developed a system that can handle multiple, concurrent connections to a defined problem.  Think of the defined problem as a project template.  The template identifies what information is sought, and the users provide information about the problem using the project template.  In other words, define the problem - then solve the problem.  To this end, we have developed a way to discretely define and present a problem - how to invalide patent claims - to a group of people.  The system can receive input about the problem and mesh that input with input from other group members.  Isn't this what crowdsourcing is all about?

A more concrete example is a joint defense group defending against patent infringement allegations.  One of the litigation defenses invariably is that the patent in question is not valid.  The project template can be used to receive information from attorneys, at different law firms.  This information is directed to the claimed features found in the prior art, to form the invalidity contentions relied upon the joint defense group members.  We see the beginnings of a crowd for this type of project.

The project template could even be used to receive prior art search result submissions.  In this case, the searchers would be the crowd.  Sounds simple enough and there are companies that already provide this service.  Now imagine that instead of commissioning a search and having attorneys review the results, the searchers themselves have access to a platform for sharing their understanding of the prior art.  Not just the prior art references, but information contained within each prior art reference as it relates to the defined problem.  Not a writeup in a document, but data entered into a database to be mixed with the data from all other searchers.  Under this approach, the individuals most knowledgeable about the prior art (the searchers) take the first pass at the analysis and record their analysis into the system.  Attorneys would then review and vet how closely the analysis matched features found in the claims.

Even Article One Partners, a noted leader in the search field, has not achieved this level of project focus.  In the Article One crowdsourcing model, a search request is broadcast to a vast team of researchers.   The researchers then review their archives and report back their findings by identifying relevant prior art, each vying for a bounty.  The knowledge and understanding that each researcher has about the prior art is not transferred in this process, however.  What if this could be done in a different way so that the searchers knowledge is transferred in the process?

We think we can fill this gap with our system.  Our system already provides a bridge that allows users to transfer their knowledge about the prior art using a project template.  The information entered into the project template can also be used to generate invalidity charts because the template fields are already linked to the claims of the target patent.  We have even considered a way to anonymize the identity of individuals inputting information into a project template if anonymous submissions are desirable.

We believe this is a different way of thinking about prior art searching and litigation workflow.  We hope you'll give us a few moments of your time and consider having us demonstrate the system to you and your team.

Four Types of Crowdsourcing - Outside In Marketing by Steve Keifer

I came across a blog post that reviews a book authored by crowdsourcing blogger Jeff Howe.  The blogger, Steve Keifer, summarizes what Howe's book describes as four variants of crowdsourcing:

In the final chapter of the book, Howe describes the four primary types of crowdsourcing – 1) Crowd Wisdom; 2) Crowd Creation; 3) Crowd Voting and 4) Crowd Funding, each of which I have outlined in more detail below.

via Four Types of Crowdsourcing « Outside In Marketing by Steve Keifer.

How can crowdsourcing be applied to patent litigation?  The application of the type of crowdsourcing would likely be dependent on the usefulness of the method to a defendant or plaintiff.

The first two types are probably the most applicable to a defendant.  For  invalidity contentions - "crowd wisdom" fits well as it relates to prior art analysis and "crowd creation" makes sense as it relates to prior art searching.

For "crowd voting" and "crowd funding," each of these may work for both plaintiffs and defendants.  As to defendants, non-related industries could target, via voting, particularly troublesome patents to pursue.  In a reexamination context, crowd funding could enable non-related industries to pool resources for a large scale reexamination effort.  On the plaintiff side, crowd funding could be used as a mechanism to fund large scale infringement pursuits.

In the next post I'll try to explain how we see our software platform positioned in the crowdsourcing space, particularly in the prior art analysis context.