CNIPA Announces Amended Patent Guidelines for Patent Applications Covering AI and Blockchain

The China National Intellectual Property Administration (CNIPA) has announced amended Patent Examination Guidelines effective February 1, 2020.  The amended guidelines add a Section 6 to Chapter IX of Part Two of the Patent Examination Guidelines, which covers provisions for the examination of invention patent applications that claim mathematical limitations and business methods.  The amendments are made for the special nature of examination of patent applications related to artificial intelligence, “Internet +”, big data, and blockchain.

The amended guidelines state that a claim as whole should be examined and an Examiner should not separate out features directed to a business or algorithmic method.  Instead, all the limitations recited in the claims should be taken as a whole, and the technical means involved, the technical problems to be solved, and the technical effects obtained should be considered.

6.1.1 Examination in accordance with Article 25 (1) (b) of the Patent Law

If a claim relates to abstract algorithms or purely commercial rules and methods, and does not contain any technical features, then the claim belongs to the rules of intellectual activity stipulated in Article 25 (1) (b)¹ of the Patent Law and the claimed method should not be granted a patent. For example, a method for establishing a mathematical model based on an abstract algorithm and not containing any technical features  should not be granted under Article 25 (1) (b) of the Patent Law. Similarly, a method of rebating as function of consumer spending does not  not contain any technical features and should not be granted under Article 25 (1) (b) of the Patent Law. 

6.1.2 Examination under Article 2 (2) of the Patent Law

If a claim passed muster under 6.1.1, the claim must also be examined under Article 2 (2)² of the Patent Law.  Again, the claims must be considered as a whole. If a claim recites technical means to use a natural law  for a technical problem to be solved, and the technical effect conforming to the natural law is obtained, the solution defined by the claim belongs to the second paragraph of Article 2 of the Patent Law and is subject matter eligible.

6.1.3 Review of Novelty and Inventiveness

All limitations of a claim should be examined for novelty and inventiveness, even those that are directed to algorithms or business methods.

6.2 Examples

The amended guidelines also provide multiple examples to better understand how the guidelines should be applied.

Method of Building Mathematical Models

A method of establishing a mathematical model increases the accuracy of modeling by increasing the number of training samples.  The example claim includes, training an initial feature extraction model; processing feature values according to the feature extraction model to get extracted values for each training sample; combining the extracted feature values and labels to get training samples; and training an initial classification model to get a target classification model.

Per the guidelines, the feature values, extracted feature values, label values, target classification models, and target feature extraction models of the processed training samples are all abstract and general data. The process of training the model is a series of abstract mathematical method steps. The final result is also an abstract general classification mathematical model. This solution is an abstract model building method. Accordingly, the claim is not subject matter eligible under Article 25 (1) (b) of the Patent Law.

Training Method of a Convolutional Neural Network (CNN)

The CNN is trained to identify images of different sizes.  The guidelines explain that the data processed in each step of the claim is image data and each step processes the image data, which reflects the neural network training algorithm and image information.  What this solution solves is how to overcome the technical problem that the CNN can only recognize images with a fixed size. The technical effect of the trained CNN model is it can recognize images of any size . Therefore, the solution for this invention patent application belongs to the technical solution specified in Article 2 (2)

Communication method and device between blockchain nodes

Before a node in a blockchain establishes a communication connection, it can determine whether to establish a communication connection according to the CA certificate carried in the communication request and a pre-configured CA trust list. This reduces the possibility of nodes leaking private data and improves the security of data stored in the blockchain.

The problem to be solved in this application is how to prevent blockchain nodes from leaking user privacy data in a blockchain network. It belongs to the technical problem of improving the security of blockchain data. By carrying the CA certificate in the communication request and pre-configuring the CA trust list, the method determines whether to establish a connection, restricts the nodes that nodes can establish a connection with, and uses technical methods that follow natural laws to obtain the technical effect of secure communication between nodes and reducing the possibility of nodes leaking private data. Therefore, the solution for this invention patent application belongs to the technical solution specified in Article 2 (2) of the Patent Law and is subject matter eligible.

Method of Rebating Based on Consumer Spend

The invention patent application proposes a method of rebate based on spending amount.  The computer executes a set of rebate rules to give the consumer coupons for future consumption, thereby increasing the consumer’s willingness to spend and increase profits for the merchant.

This solution involves a method of generated a rebate based on spending. This method is executed by a computer. The data processed is the user ’s spending data. The problem to be solved is how to promote user spending. It does not constitute a technical problem. The method claimed is a computer executing the artificial rebate rules, but the limitation on the computer is only to determine the rebate amount according to the user’s spending amount according to the specified rules. Therefore, no technical means are used. The effect obtained by this solution is only to promote user spending. Is not a technical effect. Therefore, the invention patent application does not belong to the technical solution specified in Article 2 (2) of the Patent Law, and is not subject matter eligible.

Other examples include method of enabling use of a shared bicycle (subject matter eligible); economic prosperity index calculation based on power consumption (not subject matter eligible); method of detecting a falling state of a robot using sensor data (inventive); multi-robot path planning (not inventive); logistics distribution method (inventive); and dynamic viewpoint visualization (not inventive).  The full text of amended guidelines (Chinese only) are available from CNIPA as are the 2010 version of the guidelines in English.

1 Article 25 reads, in part,

Patent rights shall not be granted for any of the following:
(1) scientific discoveries;
(2) rules and methods for intellectual activities;
(3)methods for the diagnosis or treatment of diseases;
(4) animal or plant varieties;
(5) substances obtained by means of nuclear transformation; and
(6) designs that are mainly used for marking the pattern, color or the combination of the two of prints.
The patent right may, in accordance with the provisions of this Law, be granted for the production methods of the products specified in Subparagraph (4) of the preceding paragraph.

2 Article 2(2) reads Inventions mean new technical solutions proposed for a product, a process or the improvement thereof.

  Back to All Resources