Is patent analytics the future of patent searching?

Hi "N"

Although we may use a semantic search to draw together a list of relevant publications, there will always be a need for a human in the latter stages to make a final determination. Necktop computers are far more able to distinguish semantic notions than any other type of computer. It's just that we do it much more slowly.

Hi there;

It is just a fact that the amount of documents to be searched increase with a pace that it sometimes, or even often, it does not matter how good your seach profile is, it still retrieves too many documets to be reasonably browsed necktop (nice word, like it..;0), at least for "routine" searches. The obvious remedy is to add new restrictions to the search that reduces the recall.

In view of this, alternative strategies, e.g. using semantic searching for potentially first broadening the search base, and then apply analysing tools for doing the "drill-down" (instead of adding sketchy boolean manipulations), need to be taken seriously. It may be the same kind of paradigm shift when we abandoned manual searcing if favour of online. Also then similar concerns were rised about how to gain a deeper understanding of things just from a flickering screen? But tools got better, and they continue to do so.

However, patent analytics is no magic stick. Thorough knowledge of searching, indexing and classification will never get old-fashioned. Beware of "one-click" analyses! It is also of importance to select a tool that is not too much of a "black box", but is still easy to use. The learning threshold is otherwise lurking behind the corner.

One need to use discretion of course. Finding "white holes" is indeed more difficult than to spot the black ones. I would use sanity checking to complement such searches, looking into NPL and internet. After all, if there are no products out there, then why not let it have a go? A relatively new and useful feature I've seen is the capability to compare search results, and thus also portfolios. Licensing seem to gain importance, and there are more high-profile litigation cases out there than I remember seeing before. Analytics tools are indeed making progress in the areas they were originally intended for - to show the CEO you are worth listening to!

Cheers,

Kristian

Hi "N"

I don't believe there will be any human-made machine can do what human can do in thinking, imaging and all kinds of intellectual activities. If there were such machine, it will be a paradox.

Henry

---

Muchiu (Henry) Chang, PhD. Cantab

Patent Competitive Intelligence Researcher

Monte Carlo Modeling Simulation Expert

henry.chang212@gmail.com

The need for manual micro-level analysis for patentability, infringement, validity and other types of research and analysis is not going away within our lifetimes.  I find it frustrating that even very experienced people in this field seem to think that the manual efforts will be replaced soon by text mining and data visualization tools.  All I can say is they may need a reality check.  The tools currently have extremely limited application (only to a few very specific questions high-level that arise) and they do not even touch on the large majority of questions relating to the need for and usage of patent information. 

(I did not make the original post.)

1. Patent mapping / text mining / data mining / landscape are top-down activites.

2. Prior art searches / infringement searches / validation-invalidation searches are bottom-up activities.

Neither is going away anytime soon. Having great tools definitely help.

Happy researching - Past, Present and Future.

Well, you just knew that this was going to be a lively discussion.

I like to tell people that I'm a greedy, self-centered user and that tools that make my job easier are good tools. And patent analysis tools can make part of my job easier. For me, these tools are able to crunch through raw data (from simple inventor counts to landscapes) that are really be too large or diverse for me to get through in any kind of timely fashion. Once the tools have done their job, it's time to do mine. I generally find the results from patent analytics to be the jumping off point for me to ask the best questions. Why is this grouping or segment so large? Who are these players and how are they using this technology? Why is this space so empty? Once I have my questions lined up, I start to dig into the specific patents, companies or people and see what answers I get. Not surprisingly, those so-called answers just lead me on to other questions. Also not surprisingly, I tend to use the word iterative a lot.

So, as a number of people have already stated, the human mind is still the most important tool we have. This is true not only in being able to make connections, decisions and more related to patent information, but even more so in that we do not live or operate in a vacuum. Patents and patent information need to be viewed in terms of company strategies, tactics, markets - unmet consumer needs - how unrelated technologies are developing - cost issues - and on and on.

Maybe one day we'll get to the point where creepy HAL / Dave conversations are happening, but I'm not worried about it yet........hey, would someone open up this door? Anyone?

I don't think the patent analytics software is supposed to replace "traditional" searching or what you call micro analysis of individual documents.
However, it seems that patent analytics is becoming a mandatory part of a search professional's toolkit. In that sense, it is the future (or even the present) of patent searching.
Unfortunately (and this is somewhat of a tangent for further discussion), if computers get better at semantic processing (remember Watson, the current Jeopardy champion?), computers may be able to replace us as searchers/analyzers in the not too distant future...even if they don't, on a metaphysical level, truly "understand" either the content of the documents being searched or the question asked by the end user. Just as most people use search engines to acquire general information from the internet, without consulting a human search specialist for that task.

--David Shobe