Patent searches need to be focused and yet comprehensive enough to capture relevant references to ensure nothing is missed. There are components of a search that can be added to ensure enough ground is covered to fill in the gaps. This edition will cover some of those methods.
Patent searches need to be focused and yet comprehensive enough to capture relevant references to ensure nothing is missed
Recognising a relevant reference
Oftentimes, images only tell part of the story and digging into the full-text is required
Sounds simple, and yet it can be a more detailed process to recognise a relevant reference. Relevance of course can sometimes be assessed very easily by comparing images. When these are very close or similar it can be easy to identify a relevant reference. Oftentimes, images only tell part of the story and digging into the full-text is required.
Embodiments are often the key in patentability and invalidity searches. Relying on images on or solely on the abstract text can cause relevant references to be missed. Embodiments are kind of a wish list and extra features are disclosed there that can be incredibly useful in determining whether a reference is relevant or not. Don’t ignore the full-text and embodiments.
Embodiments are often the key in patentability and invalidity searches
Write a list of features or concepts of what you are looking for and look for those in the text and images. Compare the reference and your list of features and see how the reference matches up.
For FTO searches, if keywords in the claims are relied upon too heavily when assessing a reference, a relevant reference could be missed. Searching in the full-text can reveal the use of the keyword you are looking for but it could be absent in the claims as the feature is described in more abstract terms. Try to include full-text searches of the most obvious keywords in the full-text should budget and time allow, for a more complete search.
Searching in the full-text can reveal the use of the keyword you are looking for but it could be absent in the claims as the feature is described in more abstract terms
Narrow Search
This can be done by searching CPC headings only or using a limited set of keywords. A trove of prior art can be located among JP, CN and KR references. Using IPC headings will capture this sector of prior art that should not be missed. If you do use a limited set of keywords, check if the numbers of references indexed in the CPC heading is low enough to be completed in full. Try to separate these smaller sized headings from those can contain large numbers of references.
If there are three concepts/features in a disclosure, try searching only two to broaden out the search.
A trove of prior art can be located among JP, CN and KR references
Broad Search
Searching IPCs only can also cause relevant references to be missed. Both CPC and IPCs should be searched. If IPC headings are to be searched, resist putting them all together and look for more granular headings that are closer to the definition of what you are looking for. Casting the net wide is good at some point in the search, but it shouldn’t be your only strategy and it doesn’t necessarily mean everything has been covered or inspected. Relevant references can slip through.
Casting the net wide is good at some point in the search, but it shouldn’t be your only strategy and it doesn’t necessarily mean everything has been covered or inspected
Time-Dependent Confidence
View it as wading across a river. The journey is complete when you get to the other side. If you stop halfway across the river, your journey is not complete. The prior art reference may be waiting on the side of the riverbank, so to speak
It’s easy to do some preliminary searches and conclude that nothing of relevance is out there. It could even be a few hours. Confidence rises and you could even stop the search. Stick to the process and go through the steps, and you may find something to prove that your first few hours of searching was incomplete. View it as wading across a river. The journey is complete when you get to the other side. If you stop halfway across the river, your journey is not complete. The prior art reference may be waiting on the side of the riverbank, so to speak.
Relying Solely on a Semantic Search
Again, confidence can grow quickly and certain decisions about the search can be made
Semantic searches are very good for guiding a search in terms of identifying close CPC/IPC classifications and a relevant reference can be found.
Likely more relevant references can be found in addition to those located in the semantic search through a traditional search. Again, confidence can grow quickly and certain decisions about the search can be made. For completeness and comprehensiveness, add other approaches of searching to your project.
Machine Translations
When I first started patent searching, I had a German and French dictionary close by. Then along came machine translations, which are great and save so much time, but you do need to double check and not rely solely on these. If you have access to several databases, use them to compare translations and always use Espacenet. For Japanese references J-PlatPat is very useful.
If you have access to several databases, use them to compare translations and always use Espacenet
Mechanical | Engineering | Software | Fintech | Oil & Gas | MedTech