Examination Handbook Revision Related IoT Technology

by Hirohito Katsunuma, Katsunuma International Patent Office, Japan

IoT (internet of things) related technology is being applied to various technical fields. In IoT related technology, all things are connected to the internet, and a large amount of data is collected on a server, etc. by sensors. A significant amount of data can be analyzed, structured, and learned by neural networks, thereby creating new value and services. Trained models and their trained weighting factors (parameters) of neural networks utilizing deep learning technology can be generated, which can find solutions or create new ideas from raw data. The database or the structured data generated for use by others can also be very important.

In addition to the eligibility of the “computer program” for a patent, which is not harmonized yet globally, “data”, “structured data”, “trained model”, and “trained weighting factor”, etc. of IoT related technologies are new issues in the field of the computer related inventions.

In Japan, inventions using computer software such as controls of apparatus, which utilize the law of nature as a whole, are considered to be eligible for patents. However, an invention using computer software as a whole must fulfill the condition that the information processing using computer software must be specifically implemented by hardware resources.

In response to the development of IoT related technology, the Japan Patent Office (JPO) revised the Examination Handbook on March 22, 2017, adding examples of terms treated as “programs” or “equivalent programs”, and examples of inventions of IoT related technology.

In Japan, the invention of the “computer program” was stipulated in Article 2, Paragraph 3, Item 1 of the Japanese Patent Law as an invention of a product, which was introduced by amendment in 2002. The “computer program” is defined to be a set or sets of commands to a computer combined to produce a result (Article 2, paragraph 4 of the Japanese Patent Act). However, as “data” itself is not commands for the computer, “data” is not a “computer program” as stipulated in the Japanese Patent Act. Consequently, in the past, an invention claiming “data” was considered to be a “mere presentation of information”, and could not be patented.

As the importance of the “data” or the “structured data” increased by the development of IoT related technology, the JPO decided to treat “structured data” or the “structure of data” as a “computer program equivalent” and admit its eligibility for a patent. Also, the JPO will treat a “trained model” of a neural network as a “computer program”, and consider it to be patentable.

Consequently, examples of terms which are treated as a “computer program” or a “computer program equivalent” are added to the Examination Handbook, in Annex B “Computer software related invention”. Terms such as “module”, “library”, “neural network”, “support vector machine” and “model” are listed. Claims having those terms at the ends are eligible for patents.

In the examination of IoT related inventions, the inventive step is the same with that of other technical fields in principle. However, if the invention has advantageous effects in utilizing information by connecting things to the internet, such advantageous effects will be considered to be an element positively contributing to the inventive step.

Since IoT related technology is often composed of a system in which a plurality of devices are distributed through networks, there are considerable numbers of patent applications claiming parts of IoT systems. The JPO will treat those patent applications as sub-combination inventions which are the same with those of other technical fields.