Shaping the future of
patent search AI
PQAI is a not-for-profit, collaborative initiative to build
an open-source ecosystem of AI components
to drive innovation and improve patent quality.
An initiative by AT&T
Vision
PQAI stands for Patent Quality Artificial Intelligence.
To make the patent process more transparent for all involved, significantly grow the number of inventors, thus accelerating the pace of innovation, and simultaneously improving the overall quality of patents.
Open-source AI for Patent Data Mining
Open source software (OSS) is a big driver of freedom, trust, and innovation in the digital age.
What Linux did in the OS space and what Mozilla did in the browser space are results of open source initiatives.
It’s time to bring open source to patent search.
PQAI enables researchers, developers, and tinkerers working with patent data to leverage advanced AI, get community support, and help them make big impact.
- as seen on -
Help PQAI Democratize Patent Search
Making breakthroughs possible
Developing an effective natural-language based patent search platform is one of the most biggest problems in the field of patent data mining. PQAI aims to bring the best minds working in this area to collaborate and develop AI components to tackle this huge challenge.
Bringing cutting-edge technology to users
A lot of excellent work done by researchers ends up gathering (digital) dust in some obscure Github repositories or research papers. PQAI aims to enable researchers to create big impact with their work by making it available to real users.
Speeding up the inventive process
A big fraction of patent applications filed at patent offices around the world are rejected because it is found that the inventors ended up reinventing something already patented. PQAI aims to help inventors find all the inspiration and prior work on their problem.
Improving patent quality
The decision of grant a patent is largely on prior-art search done by patent examiners, but with so much data, it is getting increasingly difficult to run thorough search. PQAI aims to provide patent professionals with tools to run accurate searches by leveraging AI.
Leveling the playing field
Open access patent tools help under-resourced inventors from diverse backgrounds and start-ups or small organizations to accurately assess their inventions. It helps them achieve more success at the patent office under budget constraints.
Driving trust in the industry
It is known that AI can have subtle biases that can make it act in unintended ways. That’s why it is important for AI programs to be developed via an inclusive and open process. An open source development model is the best way to impart trust and transparency.
Initiative Started By
PQAI Roadmap
The need for an open source initiative in the patent AI domain was recognized and the idea of PQAI was conceived
PQAI developed its first prototype of AI patent search engine, working on a specific technology (wireless communication)
Coverage was expanded to in US patents except those related to chemical and biotech domains
A new version of 102 (anticipatory) prior art engine was released that leveraged advanced AI models such as transformers
Beta version of 103 search engine was released that could identify combinations of references as prior-art
Capability for finding results similar to those shortlisted by a user added
New version of 102 prior-art search AI model, trained on bigger dataset and using a new and advanced technique rolled out
PQAI Search Tools
Whether you’re an inventor, a patent examiner, a patent attorney, or a business owner, if you benefit by deriving insights from patent data, PQAI can help you.
By leveraging an AI-first search model, PQAI makes patent data exploration easy and efficient.
Committed to security, privacy, and unbridled innovation, PQAI empowers innovators with state of the art AI tools.
Prior Art Search Engine
Uses AI to find the most relevant prior art references. It also offers a combinational prior art search to assess the possibility of getting a 103 (obviousness) rejection from the patent office. Users can export results in informative PDF reports for further analysis and discussions with other stakeholders. Also find insights on patenting trends, etc. relating to your invention.
Concept Extractor
Already found a related prior-art reference or two and want to use it as a point to find more such references? Run them through concept extractor to find important keywords and concepts useful for further search.
Art Unit Predictor
Find out the most probable USPTO group art units, where your patent application may be routed for examination. Thereafter you can study its overall statistics relating to allowance and rejection ratios, prosecution cycle and backlog to anticipate potential prosecution.
Keyword Suggester
Patents often use different words to describe ideas than the ones we use in everyday language. Find out popular terminology surrounding the concepts your invention revolves around.
While conducting the prior art searches, this tool provides valuable keywords for a more comprehensive search.
CPC Lookup
Find out the most relevant IPC/CPC codes based on a textual description of a technology. IPC/CPC codes can help you narrow down the technology area of your interest and find all the related patents filed in it easily.
Contributors
Mentors
Technical Partners
SaaS service provider
Let’s join forces!
Excited about the future of AI patent search?
Join the PQAI community of like-minded people volunteering their efforts to shape the next generation AI in patent searching. You can help by development, testing, sharing your ideas, asking questions, or simply spreading the word.
Since its beginning PQAI has been fueled by the vision and efforts of contributors, users, and fans who’ve helped us improve and develop a state-of-the-art ecosystem of AI tools for patent data exploration.
If this vision resonates with you, we are keen to hear from you.