As the use of ‘artificial intelligence’ technologies have increased in the patent field, WIPS has come to develop an AI Search to provide an easier search method for users who have difficulty writing search queries. AI Search enables quick search and review for users who are familiar with technology but have difficulties making search queries using operators, and also provides qualities and save time for experienced users who do research and analysis works.
In this issue, we’ll look into AI Search, which is highly used among R&D researchers.
It is a new search feature that breaks away from the existing Boolean search method and inputs sentences. It is a search method that selects and provides only documents with high similarity by measuring the similarity between documents by establishing an AI datdabase applied with neural network deep learning algorithm technology.
☘ Some of the features:
➤ Provides search results in order of high
similarity to the contents entered in the search box
➤ Similarity score : 1 to 99.99
➤ Search by limiting the technology field using
the WIPO’s 35 technology classification
➤ Real-time automatic translation with no need
for language conversion
(Available languages: English, Chinese, Japanese, Korean)
☘ Comparison of characteristics of general ranking search and AI search method.
AI search extracts a list of search results
through the process of inputting sentences in natural language or uploading
files > real-time translation by translation API > data analysis modeling
processing > measuring similarity in data cubes. Through tuning of the
analysis algorithm, WIPS has built its own dictionary optimized for technical
terms of the patent documents, and syntax is analyzed focusing on technical
keywords.
☘ AI Search process
Now, let's take a look what WIPS Global's AI search has;You can search by entering a patent number or texts/ phrases/ sentences rather than a search queries. You can possibly limit the search scope to specific technologies from 35 technology classifications of WIPO.
. Sentence Search : Real-time translation is proceeding
according to the input language, and the inference process is proceeding with
real-time modeling using the META information of the model. Calculate the
similarity of the document in the service model of the input text model result.
. Number Search : Documents similar to the
entered patent number are provided as a result using the search logic.
. Chart Filter : which Shows analysis charts for search results and allows you to filter each chart result.
. Result List : Provides the final list as a
result after duplication logic processing (document/patent family) and shows
the similarity measurement score accordingly.
. Re-search : Documents with high similarity to
the relevant document are extracted by re-search.
It is a block mode of search results that show
the result list together with representative drawings.
Key phrase data of the document is provided as
core keywords for each document. When the mouse over to the representative
drawing, individual drawings appear.
Increase your work efficiency with AI search.
No comments:
Post a Comment