Patent prior art search
12 record(s) found.
|Country / Territory||Institution||Business application||Description|
|Austria||The Austrian Patent Office||Patent prior art search|
In early 2018 the Austrian Patent office was trialing several commercial providers for application of AI to pre-search of patents.
|Canada||Canadian Intellectual Property Office (CIPO)||Patent prior art search|
CIPO's Patent Branch uses commercially available semantic AI search engines (Questel, STN, Clarivate Analytics) to assist in conducting searches for prior art and citations. These tools rely on machine learning algorithms to better detect linkages between citations, applications, and the current state of the art.
Patent examiners also make use of Google's algorithms, specifically within their "Translate", "Patent", and "Scholar" tools for machine translation and access to full-text documents and claims forms from contributing international patent offices in real time with citation metrics and related scholarly publications.
For data manipulation CIPO uses the Vantage Point text-mining tool for discovering knowledge in search results from patent and literature databases while providing methods to refine, automate, import, etc. the raw data produced.
|European Union||European Patent Office (EPO)||Patent prior art search|
The EPO has been active in developing business solutions using machine learning and AI for patent searches at various degrees of implementation: Automatic Search of prior art for incoming patent applications; and Automatic generation of queries.
The EPO has generated its own reference data (gold standards) and system for measuring the performance of automated search tools.
The EPO also makes use of commercial products in the automatic annotation area through software providers in different projects.
|Finland||Finnish Patent and Registration Office (PRH)||Patent prior art search|
In early 2018, PRH was testing an AI solution called Teqmine by Teqmine Analytics Oy for patent classification and prior art search. PRH's near-term aim was to compare the system to existing commercial systems (such as Innovation Q Plus) for finding documents that are similar to a given sample text.
The system finds publications that are similar to the application being analyzed by using the vocabulary and bigrams of the application. The input to the system is the text (description, claims, and abstract) of the application. Based on the frequency of the words and bigrams extracted from this input file, the system determines the activity levels of a number of topics, and determines a number of similar publications where these topics are active at similar levels. These topics were generated when the system was trained on the whole patent corpus (WO, US, and EP patent publications from the past few decades). The system runs on PRH's own server, and the server can only be accessed from within the PRH network. Therefore, the system can also be used for non-public applications. The system processes a patent application in less than two seconds.
The publications in the output file are usually broadly related to the topic of the application. Often at least a portion of the most common patent classes of the publications are related to the application in a meaningful way. However, sometimes the publications are not related to the application or invention, especially when the application uses very common words to describe the invention. The system thus cannot be relied upon to find the relevant prior art, but it may in some cases point towards a useful direction. Currently, the system does not significantly speed up the prior art search.
|Germany||German Patent and Trade Mark Office (DPMA)||Patent prior art search|
In 2016 the DPMA initiated a project for the implementation of a central service for the prior art search in various data sources of the DPMA (e.g. electronic file, specialist databases, etc.). The central service uses algorithms to improve textual similarity search.
|Japan||Japan Patent Office||Patent prior art search|
In early 2018 the JPO was validating its systems to verify possible uses for AI to support the formulation of search terms and queries when conducting prior art searches. Using text data of examined patent documents and the retrieval history of search queries used in the examinations, the JPO was validating a function to suggest keywords and patent classifications that should be included in search queries.
|Morocco||Moroccan Industrial and Commercial Property Office (OMPIC)||Patent prior art search|
Since 2011, OMPIC uses Orbite Intelligence, a commercially available AI-powered patent analytics tool to searche global patent applications by technical domain or keywords. The map-based tool was introduced to meet the needs of the Moroccan Technology and Innovation Support Center network for searching the state of the art and the precedence of patents.
|Philippines||Intellectual Property Office of the Philippines (IPOPHL)||Patent prior art search|
IPOPHL uses a third-party search engine called DTSearch for its patent search operations. Similar to all other search engines, the system has the capability to perform incremental index, fuzzy search, and other functions. Although the said system is a low-end of AI, it is more powerful than the traditional database search.
|Republic of Korea||Korean Intellectual Property Office (KIPO)||Patent prior art search|
In April 2017, KIPO established an AI agreement with the Korean Electronics and Telecommunications Research Institute (ETRI). Together they are working to build a patent knowledge base for AI learning and cooperate on research for applying their developed AI system in IP administration.
A pilot model for intelligent patent search is under development with plans to be finalized by 2019. This model enhances the quality of prior art search by moving away from keyword search to a search system based on syntax and semantics.
|Russian Federation||Federal Service for Intellectual Property (Rospatent) / Federal Institute of Industrial Property (FIPS)||Patent prior art search|
Initiatives on the implementation of AI for search were launched in 2017. In 2018 it was put into operation for the examination of applications for inventions and utility models. Patent documents similarity search function implemented in the PatSearch system is being performed through a set of AI methods and techniques in combination with the best world information search practices. Currently the similarity search function operates in the database of Russian patent documents. A neural network trained on Russian patent documents on the criteria that take into account the complexity of the examination procedure and experts' experience is used for the establishment of the Russian distributional thesaurus.
|United Kingdom||Intellectual Property Office (UKIPO)||Patent prior art search|
The UKIPO trialed Derwent Innovation, a commercially available patent search tool. The tool comprises a semantic/smart search functionality that allows large amounts of plain text (e.g. claims, description) to be used as a search input. The search tool also has the ability to search non-patent literature alongside patent documents. Further features include the ability to manually set weightings of individual search terms in order to rank results in an answer set.
|United States of America||United States Patent and Trademark Office (USPTO)||Patent prior art search|
In early 2018 the USPTO was delivering a proof of concept, Sigma, using machine learning/AI algorithms to search whole documents against a corpus of documents. For this version of Sigma, patent applications were searched against granted patents and pre-grant publications (US only).