Before filing patent applications globally, it is of great importance to know whether the pursued subject matter is patentable in the targeted jurisdiction, as filing a patent application for an unpatentable subject matter will end up in waste of time and money.
As China is emerging as a significant destination for patent filings, and that the patent system of China differs somewhat from other jurisdictions, it would be beneficial for global patent filers to have a knowledge of patent eligibility in China. In addition, recently there were some changes in the Patent Law and patent practice of China, bringing something new to the patent eligibility issue. This article is trying to provide up-to-date information on patent eligibility in China.
Principles of patent eligibility
Basically, there are 3 clauses in the Chinese Patent Law connected with patent eligibility. In Article 2 of the latest version of Chinese patent law, which will come into effect as of June 1, 2021, it is provided that “The term ‘invention’ refers to any new technical solution relating to a product, a process or an improvement thereof”. In Article 5, it is provided “No patent right shall be granted for any invention-creation that is contrary to the laws or social morality or that is detrimental to public interest. No patent will be granted for an invention based on genetic resources if the access or utilization of the said genetic resources is in violation of any law or administrative regulation“. In Article 25, it is provided that “For any of the following, no patent right shall be granted: (1) scientific discoveries; (2) rules and methods for mental activities; (3) methods for the diagnosis or for the treatment of diseases; (4) animal and plant varieties; (5) methods for nuclear transformation and substances obtained by means of nuclear transformation; and (6) the design, which is used primarily for the identification of pattern, color or the combination of the two on printed flat works. For processes used in producing products referred to in items (4) of the preceding paragraph, a patent may be granted in accordance with the provisions of this Law”. According to these three clauses in the patent law, anything that pertains to technical solution and does not fall in the exclusion range defined in Article 5 and Article 25 is patent eligible.
In practice, when examining a patent application with respect to patent eligibility, the Chinese examiner first subjects the application to a test under Article 5 and Article 25 of the patent law. In particular, for examination under Article 25, the Guidelines for Patent Examination clearly stipulates that, as long as a claim contains even one single technical feature, this claim cannot be excluded from patent edibility under Article 25. In this respect, I think it is rather strict for a Chinese examiner to raise an Article 25 rejection.
If the application passes the test, i.e., the application does not pertain to subject matters explicitly excluded from patentability in Article 5 and Article 25 of the patent law, the examiner proceeds to examination under Article 2.
In examining a patent application with respect to Article 2 of the Chinese Patent Law, i.e., whether a claimed subject matter pertains to technical solution, the examiner follows a so-called “threeelements” approach, i.e., determines whether the claimed subject matter is for solving a technical problem, adopts technical means, and achieves a technical effect. If these three criteria are all met, the examiner determines that the claimed subject matter is patent eligible.
The Guidelines for Patent Examination gives further detailed provisions on how to determine whether a claimed subject matter pertains to technical solution. According to the Guidelines, the examiner shall take the features in a claim as a whole into consideration, and if the claimed solution as a whole utilizes law of nature to solve a technical problem, and thereby achieves a technical effect that complies with law of nature, the solution defined in the claim shall be taken as a technical solution and therefore be patent eligible.
The above provisions in the Guidelines for Patent Examination may be somewhat obscure, especially in that the Guidelines does not provide a clear definition as to what is “technical”, although this term is frequently used throughout the Guidelines. In practice, the focus is put on “law of nature”, that is to say, if an element in a claim has a connection with law of nature, it may be considered “technical”. To put it in my own language, if the elements in a claim as a whole extend to the objective realm, rather than confined in the subjective realm, this claim can be considered defining a technical solution.
The above-introduced principles for judging patent eligibility generally apply to all subject matters, including inventions involving computer program, inventions involving business method, and inventions involving artificial intelligence, which is a hot topic recently in IP circle. Meanwhile, there is a chapter in the Guidelines for Patent Examination specially for inventions related with algorism and business rule, which is highlighted by a number of specific examples along with detailed analysis on their patent eligibility.
Here I particularly want to note that, in the patent examination practice in China, for patent applications involving algorism, such as those related to AI, it is very important for the algorism to be put into a specific technical application to make the invention eligible for a patent. The following two examples provided in the Guidelines for Patent Examination may help to comprehend it from positive and negative aspects respectively.
A method for constructing a mathematical model, comprising:
training an initial feature extraction model based on eigenvalues in training samples of a first classification task and eigenvalues in training samples of at least one second classification task, to obtain a target feature extraction model, wherein the second classification task is different classification tasks related to the first classification task;
processing the eigenvalues in each training sample of the first classification task respectively,
based on the target feature extraction model, to obtain eigenvalues corresponding to each training sample;
combining the extracted eigenvalues corresponding to each training sample and a label value to form an extracted training sample, for training the initial classification model to obtain a target classification model;
combining the target classification model and the target feature extraction model to form a mathematical model for the first classification task.
For this example, it is commented in the Guidelines that, this method is not applied to any specific technical scenario, and is “abstract”, therefore is not eligible for a patent under Article 25 of the Chinese Patent Law.
A method for training a convolutional neural network (CNN) model, comprising:
obtaining initial model parameters of a CNN model to be trained, wherein the initial model parameters include an initial convolution kernel of each convolution layer, an initial offset matrix of each convolution layer, an initial weight matrix of the full connection layer and an initial offset vector of the full connection layer;
acquiring a plurality of training images;
performing a convolution operation and a maximum pooling operation on each training image by using the initial convolution kernel and the initial offset matrix for each of the convolution layers, to obtain a first feature image for each training image on the each of the convolution layers;
performing a horizontal pooling operation on the first feature image of each training image on at least one convolution layer, to obtain a second feature image of each training image on each convolution layer;
determining a feature vector of each training image based on the second feature image of each training image on each convolution layer;
processing each feature vector based on the initial weight matrix and the initial offset vector, to obtain a class probability vector of each training image;
calculating a class error based on the class probability vector of each training image and the initial class of each training image; adjusting the model parameters of the CNN model to be trained based on the class error;
further adjusting the model parameters based on adjusted model parameters and the plurality of training images, until the number of iterations reaches a preset number;
taking the model parameters obtained when the number of iterations reaches the preset number as the model parameters for the trained CNN model.
For this example, it is commented in the Guidelines that, this method is applied to a specific technical field of image processing, is for solving a technical problem, adopts technical means that conform to law of nature, and achieves technical effect, therefore pertains to technical solution defined in Article 2 of the Chinese Patent Law and is patent eligible.
Interestingly, my experience is that such a method defined in the above example 2 would likely be rejected in U.S. patent examination practice, for pertaining to abstract concept. The below example 3 may further show the differences on patent eligibility between China and U.S. practice.
A method for evaluating exploitation value of a geological resource, comprising:
acquiring geological parameters of a geological structure in a plurality of designated dimensions, the geological parameters being for characterizing a state of the geological structure in the designated dimensions;
matching the geological parameters with a threshold for each of the designated dimensions, respectively, to obtain a characterization value of each of the designated dimensions;
forming a characterization vector with characterization values of the designated dimensions, wherein the threshold represents a minimum value of the geological parameters in the designated dimensions corresponding to a worthy geological resource, the worthy geological resource corresponding to the minimum value of the geological parameters belongs to the same category as the geological resource and has been exploited; and
performing an evaluation operation based on the characterization vector to obtain an evaluation value of the geological resource in the geological structure.
This claim comesfrom a real case, but has been appropriately adapted to be presented herein. The same claim has been filed with the Chinese patent office and USPTO. In examination by the Chinese patent office, the examiner has not questioned patent eligibility of this claim and the application was eventually granted. In contrast, the U.S. examiner rejected the patent application, asserting this claim pertains to abstract concept.
Seen as such, I’m of the opinion that China is somewhat flexible than U.S. in patent eligibility. This flexibility is also reflected in other aspects, among which an example is computer program-related inventions.
Similar to U.S. standard, computer program per se is not patentable in China. However, there is a workaround. Under the patent examination practice in China, even if a computer program-based invention is described purely as a software in the specification, it is still allowed to claim an apparatus, a so-called “virtual apparatus”, which is unique to China’s patent examination practice. As a cursory example, for a computer program which causes to execute a process A, a process B and a process C, it is allowed to include an apparatus claim in the claim set as long as the program is applied to a specific technical scenario, like,
An apparatus comprising
a device for process A,
a device for process B, and
a device for process C.
The Guidelines for Patent Examination makes it clear that the devices in such an apparatus claim are considered as program modules, rather than physical components.
Again, my experience is that such an apparatus claim will be considered claiming in substance a computer program per se and rejected in the U.S. patent examination practice.
The Chinese patent office is positively embracing the emerging technologies, such as artificial intelligence, internet, big data and block chain, in terms of patent eligibility, and I’m of the view that the Chinese patent office is quite friendly in this respect.
Meanwhile, when applying patent in China for an invention involving algorism, computer program, business rule or something like that, it is crucial that the invention as a whole has a “technical” attribute to make it patent eligible.