How to find similar patent families
What Is Similar Family Searching?
Whilst looking at a patent family in Cipher, the similar family search allows you to generate a list of similar patent families in just one click. This can support a range of workflows within Cipher, but primarily it allows you to expand the analysis of relevant patent families. You can also generate the searches using multiple families, to widen the similar family search.
You’re also able to use free text to find similar patents. Meaning you can, for example, copy in details of your own unpublished inventions and examine what similar patents already exist.
The purpose of this type of search is to identify other patents that may be relevant to the same technology or invention, which can be useful for patent analysis, patent landscaping, and competitive intelligence.
The video below shows you how to generate this type of report. More information can be found in the article below:
How does it work?
Cipher uses a very sophisticated proprietary patent linguistic algorithm that has been tried and tested over the past two years across our Universal Technology Taxonomy (“UTT”) classification. It is more advanced than most other systems on the market and will typically therefore provide better results than other similarity search tools available on the market.Similarity searching starts with vectorising every patent family in the universe (think of this like giving each patent family a unique fingerprint).
Each patent family can then have its vector (fingerprint) compared against others to identify vectors (patent families) that is closest to it, returning the closest results based on the chosen sample size (50, 100, 1000 etc.).
Cipher’s deep learning model (“algorithm”) is specifically designed for patent linguistic tasks and uses the patent title, abstract & claims to generate a vector for each individual patent family. It is a similar process to how the Chat GPT model operates.
How To Use Similar Family Searching:
Starting with a Patent Family:
If you do not have any relevant Cipher Family ID's, you will need to find at least one to begin your search. This can come from anywhere within Cipher, any report or upload.
-
To access the data view in Cipher, select the 'burger' three line icon in the top left of your report:
-
You can click the Cipher family ID if you'd like to learn more about that patent family (click the family ID below).
- Once you’ve found a patent you want to use, from within that family you can then navigate to the top right of the page where you’ll find a search icon (next to the Family ID - see below image). You can also go to the similar family search directly, from the cipher family ID view above:
- By default Cipher will generate 200 similar patent families, however this can be reduced or increased by changing your selection from 'number of results' dropdown (arrow 1 below):
-
Cipher ranks these in order of most similar, but you can order them in other ways such as 'by assignee'. These options are available from the 'sort by' dropdown (arrow 2 above).
-
The Cipher family ID of your patent will populate the search box (arrow 3 above).
To understand the result of the similar family search, look to the right-hand-side of the screen where you will see the similarity score displayed. This indicates out of 100, the likelihood that this patent is relevant to the input data. The higher the number, the greater the similarity.
Adding more patent families to your similarity search:
-
If you wanted to add more families to the search you can simply copy them into the Family IDs search box (arrow 3 above). There is no limit to how many you include and the more relevant families you add, the better. Once you have all the numbers in, simply select how many results you want and run the search.
Starting with a Cipher Family ID:
If you already have the Cipher Family number(s) you want to start with, you can jump to the Similar Family Search page from Cipher's homepage.
Either type the family ID into the Search Cipher box (below) and click the magnifying glass icon:
or click on the link in red below the search box that reads 'similar families'. This is the best option if you have multiple families that you want to include in your search:
Searching with Free Text
Similar family searching also allows for inputting free text:
This is not a boolean search and will not look for exact matches, but instead text with similar content.
The text itself can come from anywhere, and you can copy in as much as you like. A typical workflow may be to evaluate invention disclosures or unpublished patents from within your company. Naturally patents in this stage of their lifecycle aren’t searchable so the text allows you to get ahead and scan what already exists (In exactly the same way as the patent numbers, simply input the text, select how many results you want and then click search.
Can I combine the similarity search with free text?
Yes absolutely, the more relevant information you can provide the tool will only enhance your results. This includes the combination of Cipher family ID’s and free text.
How does Similar Family Searching differ from Semantic Search?
Here are the key differences to semantic searching/boolean searching in Cipher:
-
The results are much better (more relevant results, more results found)
-
You can search based on patent numbers
-
You can use whole paragraphs of text
-
The search is specific to patent text, not 'general text similarity'
-
You can combine multiple inputs into one search for example, patents, blocks of text, and technology names
Exporting Results & Creating Reports from the data
Once you’re happy with the size and relevance of the pool of similar patents. You have the option to export these into excel, or you can push the families into a new cipher report, where you’re able to analyse them across the classic Cipher visualisations and metrics.
Before exporting to excel it is crucial that you pick the columns of information that matter most to you (exactly the same as the usual Cipher export). To do this head to the column’s dropdown: Here you will see the default selections already ticked. Tick and untick, when you’re happy and ready to export: Click the options and hit download into csv.
The option to push the data into a new report lives in the same dropdown. Again, when you’re happy with the scope and size of the patent list, simply select ‘build report from this list’. This will automatically load a report with every family in your list.
With any questions about similar family searching, please contact a member of the Cipher Team.
Need help? Request support and we'll connect you with an expert.