Modern methods for improving conventional prior-art searches

An efficient prior-art search, using traditional means, takes time, effort and a lot of expertise. Nowadays, the focus is continuously shifting toward a combination of elements that facilitate the retrieval of results not only with efficiency but also with remarkable speed. As precision is an uncompromising element of any prior-art search, it is important to incorporate modern searching approaches. To this end, cutting-edge AI-based technology tends to move further away from the classic searching method and helps to look at patent data from a different angle.

When agility is key

In this case, semantic search tools are the answer. Most software products in this category are, to some extent, AI-based, and will use some form of a machine learning model trained on patent data to perform complicated calculations and processes such as keyword extraction, fingerprint generation and pattern analysis. Of course, when it comes to the precision of these models, the devil lies in the details. If correctly set up and fine-tuned, these AI-based software solutions can identify the most similar prior-art in a matter of seconds. For these reasons, many organizations and professionals already benefit from semantics in their pursuit of quickly understanding the existing patent landscape.

Landscape overview and added quality

AI-assisted techniques are the fastest and most efficient choice for patent landscaping. But can a semantic tool make a difference for those who prefer to use Boolean searching to identify prior-art?

Let us start by having a look at one of the most commonly used methods – the classification search. An expert would build this type of database query using IPC (International Patent Classification) or CPC (Cooperative Patent Classification) codes, which were introduced to retrieve prior-art more productively. Through these classification systems, the documents are organized in different groups according to technical content.

Back in the day, patent information used to be kept on paper, but with the dawn of the electronic age came the possibility to save these files digitally. However, this did not render classifications useless, quite the contrary. Even when the information was stored electronically, it still proved useful for searchers to be able to query prior-art databases in language-independent manners. This approach was seen as extremely valuable, and it was legally defined in 1971, through the Strasbourg Agreement, to classify patent documents under the IPC. Years later, in 2013, out of a desire to harmonize their existing classification systems, CPC was introduced by the EPO and the USPTO. The additional levels of detail emphasized the CPC's main benefits compared to the IPC and the fact that this new system could be updated more often. These are only some of the reasons why many professionals are fond of using classic searching processes.

Reduce the risk of missing relevant prior-art

To help individuals and organizations successfully apply conventional methods in an ever-changing landscape, Dennemeyer offers Octimine, a powerful text-based search tool that allows users to combine the best of both worlds. Thanks to its semantic capabilities, this software solution can add quality to a classification search in just a few minutes. After running an IPC or CPC query, the searcher may narrow down the relevant prior-art to a few essential publications. However, in some cases, vital documents can be missed, and this can happen because of the way classifications are defined and documents are filed.

As the World Intellectual Property Organization's IPC guide explains, "the expressions function-oriented place and application-oriented place cannot always be regarded as absolute." The same guide, under section VIII Principles of the Classification, states that "The technical subjects of inventions dealt with, in patent documents concern either the intrinsic nature or function of a thing or the way a thing is used or applied." Let us focus on the second part of this phrase, where the IPC guide tells us how documents can be classified depending on the field of application. For example, many technologies, such as lasers or LiDAR systems, are applicable in different industries and may be listed under a series of classification symbols. It can also occur that patent filings are misclassified either by a fault or on purpose. The latter usually occurs when an applicant wants to hide his new filings from the competition by using a classification that is less used or likely less observed by the competition.

Depending on how the traditional classification query is defined, some of these IPC or CPC codes can be left out, leading to certain relevant documents not being considered as prior-art. Here is where Octimine can step in and act as a quality assurance tool: The publications identified through classification search can be used as search input in Octimine to determine if any similar patents were missed. Such a scenario is possible because some inventions could have been classified under codes that were not taken into account during the initial search. Regardless of the field of application and the symbols used to define the query, Octimine will focus its algorithms on the full text, thus enabling the user to retrieve all relevant prior art. A dedicated feature will then direct the searcher to any highly similar patents associated with atypical classifications.

As we have seen, traditional searches can adapt by incorporating semantic tools to cover all the bases and ultimately reduce the legal risk of missing out on substantial prior-art. There are many more reasons to choose Octimine, and if you would like to learn more, please contact our team of Software Consultants.