Automated Patent Landscaping

Automated Patent Landscaping can be defined as the process of identifying all the patents related to a particular topic. It combines machine learning and law, as software and algorithms are used to perform the process. It is becoming more and more useful in conducting  Patent Landscaping because it eliminates many of the limitations posed by manually conducting such research.

This article aims to discuss the concept of Automated Patent Landscaping in detail and also the reason why it was felt that such a technology was needed in this domain. The article also includes the benefits and shortcomings of such technology.

Introduction

Patents are a form of Intellectual Property (IP) that provides a right to the owner of such patents to exclude others from operating in that area. A granted patent represents a right to exclude others from making, using, or selling the invention in the specified jurisdiction, for a specified number of years.

With the advancement in technology and innovations, there has been a rise in the demand for patent protection, in almost all fields. Therefore, a person needs to know the innovations that are taking place in the field that he operates in so that he can evaluate the areas that are still unexplored and invest more in research and development in those fields. Patent Landscaping helps in such evaluation as it identifies all the patents related to a particular topic. This is a human-driven process and it heavily relies on intricate patent data, thus making it a difficult and complex process.

Automated Patent Landscaping, represents the incorporation of Artificial Intelligence (AI) in this domain. It makes use of machine learning to develop high-quality Patent Landscapes with minimal efforts. It requires lesser human effort as technological tools are used for the landscaping. Thus, it also eliminates the chances of human error. 

Patent Landscaping

Patent landscaping is a type of research process that creates an overview of the patents that are pending or in place in a particular area. It provides the information on what are the innovations that are already patented and which of them are awaiting patent approval, in a particular field or industry. It assists the government, companies, investors, etc in analyzing the likelihood of getting the patent that is sought, and also helps them assess the risk associated with it. As Patent Landscaping provides a summary of all the patents in a particular field, thus, the stakeholders can evaluate the areas where fewer patents are claimed and spend more on research and development in that particular field.

Patent Landscaping is a human-driven process, thus it involves a tedious task of conducting in-depth research about the patents, manually. It potentially entails an expert review of thousands of complex documents and fine assessments on their legal and technical content. Moreover, keeping the landscape up-to-date for continuing reference can be just as resource intensive as its initial development. Therefore, to conduct Patent Landscaping, proper training in legal and technical aspects is also required.

It should also be noted here that Patent Landscaping is different from prior art searches. Although both involve searching for patents, they differ in use and scale. Prior art searching is focused on identifying a small number of patents, or other publications that are related to a ‘specific invention’. Landscaping is focused on identifying a large number of patents that relate to a ‘topic’ or a particular ‘field’. 

Patent Landscaping Reports

According to the World Intellectual Property (WIPO), Patent landscape reports (PLRs) provide a snapshot of the patent situation of a specific technology, either within a given country or region or globally. They can inform policy discussions, strategic research planning, or technology transfer. They may also be used to analyze the validity of patents based on data about their legal status.

A patent landscape report generally contains information about whether the competitors are patenting in the same business or research areas as one’s own company. It may also turn up information on any new entrants to the market in that area and whether there are any patents pending that are similar to one’s outstanding or upcoming patent applications. A patent landscape is not an exhaustive list of patents. Instead, it contains a snapshot of the patent situation at a particular time.

Automated Patent Landscaping (APL)

Automated Patent Landscaping combines Artificial Intelligence (AI) and Patent Landscaping. AI can be defined as a performance by machines of the tasks that are generally associated with intelligence. Thus, computer programmes can be used to solve problems and perform complex tasks that require the application of intelligence. AI generally mitigates the errors of human performance and leads to better outcomes. It is increasingly being used in law-related professions, to perform tasks that were earlier performed by entry-level lawyers. It is mainly because of the advantages that the use of such technologies offer, like less error rate, less time consuming, working without breaks, to name a few. 

Automated Patent Landscaping can be defined as a semi-supervised machine learning approach for Patent Landscaping. It is the use of algorithms and machine learning to perform the task of an expert on Patent Landscaping. Thus, a machine can be used to perform this otherwise challenging function that can take substantial time and effort.

This approach takes a human-curated set of seed patents and expands it to populate a patent landscape. A human-curated seed set is a sound starting point as it provides human insight as to the contours of the landscape. The seed set is then expanded using citations and class codes. The initial results are often over-inclusive, but this is mitigated by pruning out less relevant patents using a machine learning model. The model provides a form of double verification.

Need for Automated Patent Landscaping

Patent Landscaping is a complex procedure and requires a lot of time and effort of the person performing such a task as it involves a load of data that has to be gone through. It requires searching through a pool of data and that too manually. Therefore, there are also high chances of human error. With the introduction of AI in this process, Patent Landscaping has become less time consuming and there has been a significant decline in the effort and cost involved in the process. In addition to that, Automated Patent Landscaping also involves minimal human effort, thus eliminating the chances of human errors.

Advantages of APL

  • It significantly reduces the time and effort required in Patent Landscaping. It delivers the data quickly, by entering only a few search words. 
  • These tools are best for situations that require a general overview and when lower accuracy is sufficient. For example, if a scientist wants to understand what the key technologies are in a new research domain and needs a quick overview of the most important players in that field, automated patent landscape tools provide a fast way to see this information.
  • It helps in better allocation of resources within the research efforts because by providing the basic outline of patents that are existing or pending in a particular field, it provides an overview of the areas of research where there is little threat of competition.
  • It also helps the business development teams identify the owners of the relevant technology. This is useful for considering potential merger/acquisition partners.
  • It can also be used to apply for licensing in particular technologies as it provides an overview of the key Patent holders in the area.
  • A patent landscape can provide companies with information about how likely they are to get a patent approved.
  • The is approach is quite flexible and can be adapted to new advances and future research.

Limitations of APL

  • A Patent Landscape cannot provide information about the other barriers to research that might operate in a particular area.
  • The information is machine-generated, therefore it provides a pool of information, some of which might not be relevant.
  • They can miss relevant documents and automated tags that explain the details of the automated set. Because automated tools use an algorithm to label the documents instead of manual tagging, the tags usually don’t make sense on their own and can lead to inaccurate statistics and charts with blurry results.
  • The Automated Patent Landscape tools usually only search the published patents, missing relevant documents such as research literature and business news. 
  • Automated patent landscape tools usually rely on general templates that show information in a limited manner. Therefore, these templates may be more suitable for users looking for a general overview of the Patent situation in an area, and not for the users who want to take important strategic decisions.

Conclusion

Patent Landscaping has traditionally been a time consuming and complex process relying on the careful construction of queries to identify relevant patents. Recent machine learning advances promise to reduce these costs by automating landscaping while providing scalability and accuracy. Automated Patent Landscaping has proved to be a better approach than manually doing the work. It has reduced the work of users significantly and saves their time. It provides a basis for understanding overall trends as well as risks and opportunities in a given area of technology. They provide valuable business insights, revealing information such as the most dominant players in the technology sector and what solutions they provide, with just a few clicks. Thus, this automated patent landscaping approach provides a repeatable, and accurate way to generate patent landscapes with minimal human effort.

References

Leave a Reply

Your email address will not be published. Required fields are marked *