What are the best practices for writing patent claims for software and algorithms?

1 What are patent claims?

Patent claims are the most important part of a patent application. They define the scope and boundaries of your invention and what you are legally entitled to exclude others from making, using, or selling. Patent claims are written in a specific format and language, using terms that are clear, concise, and supported by the description and drawings. Patent claims can be either independent or dependent. Independent claims stand alone and describe the essential features of your invention. Dependent claims refer to and further limit an independent claim by adding additional features or details.

2 How are patent claims for software and algorithms different?

Patent claims for software and algorithms are different from other types of patent claims because they deal with abstract ideas, mathematical formulas, and computer-implemented methods. These are not considered patentable subject matter in many jurisdictions, unless they meet certain criteria. For example, in the US, software and algorithms must have a practical application, produce a technical effect, or improve the functioning of a computer or a machine. In the EU, software and algorithms must have a technical character, contribute to the state of the art, and solve a technical problem. Therefore, patent claims for software and algorithms must be carefully drafted to demonstrate how they meet these requirements and avoid being rejected as ineligible or obvious.

3 What are some tips for writing patent claims for software and algorithms?

One of the tips for writing patent claims for software and algorithms is to use a hybrid approach that combines different types of claims. For example, you can use a method claim that describes the steps or actions performed by your software or algorithm, a system claim that describes the components or devices involved in your software or algorithm, and a product claim that describes the tangible result or output of your software or algorithm. This way, you can cover different aspects and perspectives of your invention and increase your chances of obtaining a valid and enforceable patent.

Another tip is to use clear and consistent terminology that defines your software or algorithm and distinguishes it from the prior art. You should avoid using vague or generic terms that could be interpreted in different ways or encompass existing solutions. You should also avoid using terms that are too narrow or specific that could limit the scope of your invention or exclude potential embodiments. You should use terms that are well-known in your field or that you have defined and explained in your description and drawings.

A third tip is to include technical details and features that show how your software or algorithm solves a problem, achieves a result, or improves a function. You should avoid merely reciting the abstract idea or the mathematical formula without showing how it is applied or implemented. You should also avoid adding unnecessary or obvious features that do not add any inventive value or that could be easily modified or replaced by a skilled person. You should focus on the novel and inventive aspects of your software or algorithm and how they provide a technical advantage or benefit.

4 What are some examples of patent claims for software and algorithms?

To illustrate the tips above, here are some examples of patent claims for software and algorithms from different fields and domains. Note that these are simplified versions of actual patent claims and are not intended to be legal advice or guidance. For example, one method for detecting faces in an image involves receiving an input image, applying a convolutional neural network to generate a feature map, applying a non-maximum suppression algorithm to identify regions of interest, applying a face classifier to each region of interest to determine if it contains a face, and outputting regions of interest that contain faces as the output image. Another system for generating personalized music playlists comprises a user interface for receiving user preferences and feedback, a music database for storing music tracks and metadata, a music recommender for accessing the music database and user preferences/feedback and applying collaborative filtering to generate a list of music tracks, and a music player for playing the tracks from the list. Finally, instructions on a computer-readable medium can be used to encrypt data by generating public/private keys based on homomorphic encryption scheme; receiving data set; encrypting data set using public key; transmitting encrypted data set; receiving encrypted result from server; and decrypting encrypted result using private key to obtain original result.

5 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?