Contact us

Classifying patents using machine learning

How can you know what patents are relevant to you?

Whether you are trying to understand the competitive landscape, benchmark against it, review your portfolio, explore monetisation options or perform due diligence. It’s unrealistic to be able to read them all to find out. Cipher can help.

The Big Data Problem in Patents

Data is growing.

IBM says that 90% of the worlds data has been created in the last 2 years

There are 89m patents in the world

In 2019 3.8m new patents were filed, up 35% from 2016


Why Cipher?

Because manually classifying patents is too painful.

Cipher was created to automate the manual process of sorting, tagging and classifying patents to your view of the world.

Why Cipher

<img decoding=

Cipher Classification

With Cipher’s Custom Taxonomies, you are able to see the world through your own technology lens.

View your Custom Taxonomy


What is Machine Learning?

Machine learning is a way to achieve human-like results, by training algorithms how to perform complex tasks instead of explicitly coding them.

Gmail have stopped 99% of spam through Machine Learning.

Machine Learning is everywhere today

How can you be sure Machine Learning works to classify patents?


How can we trust the machine?


How can we see if the

Cipher Machine Learning algorithm

performs when classifying patents?


…We need to test it.

Read the full academic study

How to robustly test the Cipher Machine Learning algorithm

This is what we did to test the Cipher algorithm:

1. The patents from the data sets were manually sorted

2. A small section of the data sets were used to train the Cipher algorithm, then

3. The results delivered by the Cipher algorithm were compared against the manual results


The following criteria had to be met to deliver a robust test:

<img decoding=

Accuracy Webinar

Cipher hosted a webinar on Developing a Gold Standards for testing the accuracy of the Cipher Machine Learning Algorithm to classify patents in collaboration with Patinformatics LLC. Joined by the Managing Director, Tony Trippe, the webinar covered the following points:

  • How a Gold Standard helps test the accuracy of machine learning
  • The approach to building a Gold Standard that can be trusted
  • Testing Cipher against the Gold Standard

Watch the webinar now

Insights & Events

Measuring the accuracy of AI for classifying patents – what’s the Gold Standard?

View our webinar where we discuss the Gold Standard for testing the accuracy of patent classification using AI and machine learning.
Read more

Testing for accuracy using independently created Gold Standards

Cipher scores a 98% accuracy in an independently created test.
Read more

The role of AI in evidence-based strategic IP decisions

As patents increasingly take shape as a functional asset class, the use of artificial intelligence looks…
Read more

AI throws light on Automotive industry’s billion dollar intellectual property

Automotive Patent Trends is the definitive guide to who’s protecting what across electrification, autonomy and connectivity.
Read more

Case Studies

ARM on classifiers and managing patent budgets

Rob Calico, VP Litigation and IP, discusses the challenges of managing a sophisticated portfolio in a fast-moving technology environment, and how Cipher has helped them manage their patenting strategy and budgets.
Read more

Richardson Oliver Insights on the challenges of IP Teams

COO of Richardson Oliver Insights, Erik Oliver, talks about the challenges faced by IP teams with managing high value portfolios and the increased demand for both better data and better ways to communicate the value of patents.
Read more

Centrica on the value of Cipher

Charles Clark, Director of Intellectual Property at Centrica, discusses the efficiency gains from using Cipher and the tasty bits of data gleaned from using Cipher’s AI over conventional search.
Read more

With Cipher you can…

Portfolio Optimisation connecting cogs

Optimise your portfolio

Ensure you have the right portfolio to meet your strategic patenting objectives.

Read more
Competitive Intelligence

Gather competitor intelligence

Understand who’s doing what by automating patent to technology mapping.

Read more
Cross Licensing

Model cross licensing

Combine patent and revenue data to determine rational licensing outcomes.

Read more

Manage your budget

Justify patent budgets and communicate the impact of your investment.

Read more

Conduct due diligence

Automate manual reviews for efficient execution of M&A and licensing transactions.

Read more

Tackle inbound patent assertion

Be prepared with evidence to create a fast and effective threat assessment.

Read more

Benchmark your portfolio

Assess your portfolio in comparison to other owners through your technology lens.

Read more

Monetise your portfolio

Identify opportunities to create value through licensing or sale of patent assets.

Read more
Technology Trends icon

Predict technology trends

Track new technologies and discover the unknown owners of future innovation.

Read more
Icon of a meter

Create Risk Mitigation Strategies

Understand the materiality of your threats to define your risk mitigation strategy.

Read more

Improve your patent strategy now

Speak to the Cipher team today.

Review Your Portfolio