Case study - How to find hidden patents for a patentability search using trial version of Ambercite AI

Ambercite AI is an advanced AI method of finding similar patents, that due to its unique algorithms can find 'hidden' patents missed by conventional patent searching.

Recently we have released a trial version of Ambercite AI, which dispite some limitations can still be used as a fully functional and rapid patent search tool. Applications for Ambericite AI include patent invalidation and patent examination - and this case study will even show what even the trial can do for a patentability case study.

The archetypal invention is often cited decribed as a mousetrap, and these days everything seems to be getting networked, so maybe a mousetrap that connects to your phone sounds like a good idea. But has it  been done before?

Ambercite can find very similar patents to one or more starting patents - in the case of a patentability search, we often need to determine the starting patents. Google patent is a very popular free search engine, so I tried the following query in Google Patent, and looked at the first few patents:

Mouse trap query.gif

This was not so successful as a query - 'mouse' has a range of meanings within the IT world. Maybe we should try this query based on rodents instead:

Rodent trap query.gif

This was more successful - the top four patents all appear to be relevant. But not the fifth, or indeed the remaining results, at least on the first two pages. But is this list of patents complete? What other relevant patents are there? - that might 'hidden' due to unexpected keywords or class codes?

>>>>>>>>>>>>>>

There are a few different ways of asking this question. We could try different combinations of keywords. We could get more sophisticated, log into a subscription based patent search engine, determine some useful keywords and class codes, and then spend time going through all of the mostly irrelevant patents this approach will uncover.

Or - you could simply add the results you have already found into the trial version of Ambercite. 

The trial version of Ambercite allows up to 3 starting patents (the full version allows up to 200 starting patents). There are 4 patents listed above, but two are from the same applicant, so maybe we should use one patent number from each of the three applicants. The resulting query will look like this in the trial version:

Mouse trap Ambercite.gif

The top four results look like this

Mouse trap top three results.gif

While a few of these results are subscriber only, most are not. We can see the first patent, US9015987 appears to be a rodent trap with smartphone monitoring. And so in fact are other other patents in this top 25 list, including:

  • US8599026B2Animal control system
  • US7530195B2Electrocution animal trap with a sender
  • US2014085100A1Wireless notification system and method for electronic rodent traps
  • US2017215407A1, Pest trap monitor
  • WO2017011916A1, A pest control monitoring system

And while not strictly relevant, there is even a patent for a bear trap - which apart from the relative sizes of bears, is still quite relevant.

Bear trap patent.gif

And we could keep going if we wish. For example, we could put the top three listed found patents back into Ambercite: 

(we don't normally recommended including patent titles along with patent numbers - but in this query Ambercite still found similar patents to the patent numbers listed)

second query.gif

And this bought up further similar patents, including

  • US6445301B1, Electronic pest monitoring system and method
  • US7026942B2, Method for signalling the presence of prey in traps for vermin and device for carrying out this method

We could keep going and try other combinations of patents, but hopefully this is enough to demonstrate the principle.

 

And in the full paid version?

The full version of Ambercite Ai allows queries with up to 200 starting patents, and up to 2000 results, with no results hidden. Ideally everyone would have access to this, but in absence of this access, the trial version can still be very useful.

 

Do you wish to find your own 'hidden' patents?

Ambercite now offers an online trial version of Ambercite Ai. Further details are found at the link below, but essentially this involves registering online (no payment details required) and then using Ambercite for yourself.

 

 

Mike Lloyd