Plain Language Statement
Project: Patent Claim Scope
Research Team
Project contact: Dr Mark Summerfield
Email: m.summerfield@student.unimelb.edu.au
Responsible researcher: Professor Andrew Christie – email: a.christie@unimelb.edu.au
Other researcher: Professor Tim Baldwin – email: tbaldwin@unimelb.edu.au
Introduction
Thank you for your interest in participating in this research project. The following will provide you with further information about the project, so that you can decide if you would like to take part in this research.
Please take the time to read this information carefully. We welcome questions about this statement.
What is this research about?
This project seeks to address questions relating to the performance of the Australian patent system. More particularly, the project aims to determine how the scope of Australian patents may have changed over time, and how it compares with the scope of patents granted in other jurisdictions. In plain terms, the ‘scope’ of a patent, as defined for the purposes of this project, refers to the technical extent of the rights granted to the patentee or, equivalently, to the extent of the patentee’s power to restrict the activities of competitors.
The stated object of the Australian Patents Act 1990 is ‘to provide a patent system in Australia that promotes economic wellbeing through technological innovation and the transfer and dissemination of technology.’ This object is achieved by granting exclusive rights to innovative individuals and companies for the commercial exploitation of inventions that they develop. The rights that are granted are limited in duration and in technical scope. These limitations operate to balance the interests of producers and owners of technology with those of users of the technology, and of the public more generally.
Broader patents, i.e. patents representing rights having wider technical scope, impose greater restrictions on competition in markets for the patented technologies. Therefore, the scope of granted patents is a key indicator of the effectiveness with which the patent system strikes the intended balance between the interests of different stakeholders.
However, measuring patent scope is challenging because it is defined using patent claims, which employ natural language to set out the boundaries of the patent monopoly. One objective of this research project is to develop metrics of patent scope by applying computational models, including artificial intelligence (AI) natural language processing (NLP) technologies, to claim text. The usefulness of these metrics will be demonstrated through a comparative analysis of the scope of patents granted under differing legislative regimes, enabling the relative scope of patents granted under Australian law to be evaluated over time, and in comparison to other jurisdictions.
The development of AI NLP models requires the availability of 'ground truth' datasets. Such datasets comprise representative samples of source data that have been evaluated, or ‘labelled’, by human experts. The datasets can be used to train models, i.e. to enable the models to ‘learn’ to reproduce the evaluations of the human experts. They can also be used to validate models, i.e. to compare model outputs with the expert evaluations to assess model accuracy.
The aspect of the project in which you are being invited to participate is directed to ‘crowdsourcing’ labelled data – i.e. collecting information through your contributions and those of other people with relevant expertise – to build a suitable ‘ground truth’ dataset.
The research project is being conducted towards the award of the degree of Doctor of Philosophy at the Melbourne Law School by Dr Mark Summerfield, under the supervision of Professors Andrew Christie (Melbourne Law School) and Tim Baldwin (Laureate Professor, Computing and Information Systems, the University of Melbourne).
What qualifications do I need to participate in this research?
We are presently seeking volunteers with experience in reading, comprehending and evaluating patent claims drafted in formats that are common in mechanical, electrical, electronic and related fields across major English-language jurisdictions. (Note that the present phase of the research does not encompass chemical, pharmaceutical and biotechnology fields.) The largest body of available granted patent claim text is published by the US Patent and Trademark Office (USPTO), and therefore familiarity with US claim drafting practice is highly desirable. We are particularly seeking qualified patent practitioners from Australia, the US, New Zealand, and Canada, along with UK and European practitioners having experience across jurisdictions including the US, to participate in the project.
What will I be asked to do?
The data-gathering exercise is entirely web-based and involves interacting with an online claim scope evaluation application.
If you choose to participate, you will register through the website. Registration will require you to provide your name and email address, and to create a password for access to the claim scope evaluation application. To complete your registration, you will need to click a confirmation link in an email that will be sent to your registered email address.
Once you have completed your registration, you will be required to complete a profile identifying:
- the field(s) of technology in which you have relevant experience in evaluating patent claims;
- the jurisdiction(s) in which you have experience in evaluating patent claims; and
- any jurisdiction(s) in which you are formally qualified, registered, and/or certified as a patent practitioner (e.g. patent attorney or patent lawyer).
You will not be required to provide any other personal or contact information.
During the exercise itself, the claim scope evaluation application will present you with a series of screens comprising pairs of patent claims sampled from one of your nominated fields of technology. You will be asked to evaluate which of the two claims presented on each screen is, in your assessment, broader in scope. The evaluation is primarily qualitative rather than quantitative. You will not, for example, be asked to assign a numerical value to the difference in scope.
There is no limit on the number of evaluations you may perform, other than the time period during which the online application is available and open for data collection. You can return as many times as you like using your registered email address and password to log in. On each visit you may perform as many, or as few, evaluations as you wish.
You may cease your participation at any time. If you simply stop participating, the evaluations that you have performed to date will be processed along with those of other participants in building the final dataset.
Additionally, you can choose to formally withdraw your participation at any time during the first 7 days after initial registration. A link is available to you on the website for this purpose during the relevant period. If you elect to withdraw, your account will be locked (i.e. you will no longer be able to log in and access the application), any evaluations that you have performed prior to withdrawal will be excluded from the final dataset, and your personal identifying information will be removed from the project records.
What are the anticipated benefits of my contribution?
Your evaluations will be used to build a novel dataset of claim scope comparisons that will be used to train and validate new computational claim scope models, including AI NLP models, capable of comparing and ranking the relative scope of patent claims not seen in the training dataset. These models will, in turn, be used to identify changes and differences in claim scope over time, and between comparable jurisdictions. This will enable large volumes of published patent data to be analysed to address important legal, economic and social question relating to the performance of the patent system.
We also plan to publish the dataset to enable other researchers to replicate our work, and to conduct their own research using the data. Making high-quality, ethically sourced data open to the research community is important for the advancement of knowledge, the development of new applications, and to avoid duplication of the time and labour invested in generating such data.
What are the possible risks?
The potential risks to you are minimal. The only personal information that will be requested from you is an email address, and details of the jurisdiction(s) and technical field(s) in which you have relevant expertise. For most participants, we anticipate that this is information that is already publicly available in online personal and/or professional profiles. The web server and application are built using open-source software platforms and frameworks that are employed by millions of websites around the world, and which include appropriate security mechanisms. In this way, all reasonable efforts have been made to protect your information from unauthorised access. As is standard practice, the password that you provide will be stored in a scrambled (‘hashed’) form and will never be transmitted or stored in plain text. Nonetheless, the usual recommendation not to reuse the same password across multiple websites and accounts applies. We recommend the use of a password manager to generate and store a strong, unique password that you do not have to remember for yourself.
Will I hear about the results of this project?
When completing your profile through the website you will have the option to sign up for email updates on the project. If you elect to sign up, you will receive email notifications of significant public outcomes of the project. Such outcomes may include the publication of reports, articles in academic journals, online articles, and the final dataset.
What will happen to my information?
The records from this project will be stored digitally on protected University systems. Your personal identifying information will be stored securely for as long as it is of research value to the project or projects closely related to the original project in accordance with the University’s storage policies, for a period of at least five years. After that time the data, other than the de-identified final published dataset, will be destroyed.
The final published dataset will not include your personal details, or any other identifying information. It is entirely up to you whether you disclose your participation to any other person or organisation.
By registering through the website, you acknowledge that you have read and understood this information. Furthermore, by participating in the project and providing claim scope evaluations, you will be taken to have consented to the use, storage and publication of the data that you have provided, as described above. However, if you formally withdraw from participation within the relevant time, this consent will be taken to have been withdrawn.
Where can I get further information?
If you would like more information about the project, please contact the researchers: Mark Summerfield (project contact) at m.summerfield@student.unimelb.edu.au.
Who can I contact if I have any concerns about the project?
This project has human research ethics approval from The University of Melbourne, ethics ID number 30794. If you have any concerns or complaints about the conduct of this research project that you do not wish to discuss with the research team, you should contact the Research Integrity Administrator, Office of Research Ethics and Integrity, University of Melbourne, VIC 3010, via telephone on +61 3 8344 1376 or email research-integrity@unimelb.edu.au. All complaints will be treated confidentially. In any correspondence, please provide the name of the research team and/or the name or ethics ID number of the research project.