In replying to the patent examination opinions, we often see such a review opinion. "The technical solution protected by claim 1 only deals with the rules and methods of intellectual activity, and is the case of Article 25, paragraph 1 (2) of the Patent Law. A patent that is not a patent protection object cannot be granted."

The Patent Examination Guide (2010) stipulates that “intellectual activity” refers to the thinking movement of human beings. It originates from human thinking, produces abstract results through reasoning, analysis and judgment, or must pass human thinking movement as a medium. Indirectly acting on natural production results. The rules and methods of intellectual activity are the rules and methods that guide people to think, express, judge, and remember. In summary, we can see that the rules and methods of so-called intellectual activities are obtained through people's logical reasoning and analysis. There is no technical field for specific applications, no technical means or natural laws, and no technical problems are solved. And produce technical effects, and thus do not constitute a complete technical solution.

So how can we avoid the situation as stipulated in Article 25, paragraph 1 (2) of the Patent Law? The author mainly elaborates from two periods:

First, when the new application is written

Agents often encounter many patents for optimization algorithms when they apply for new applications, such as optimization of clustering algorithms, optimization of rice algorithms, and optimization of lightning search algorithms. When the inventor provides the technical disclosure, the entire optimization algorithm does not give the technical field, technical scope or application engineering of the application. In order to avoid the provisions of Article 25, paragraph 1 (2) of the Patent Law in the later review of the opinion, the agent should try to avoid the situation as stipulated in Article 25, paragraph 1 (2) of the Patent Law.

1, start from the topic

When writing the title of the topic, the agent can formulate the title of the topic "a domain or engineering optimization method based on...", it is best not to write the optimization method of the algorithm, or write "a method of optimization of the algorithm" The optimization method of the algorithm is applied to the ... field or engineering, and the above two methods can reduce the probability that the invention belongs to the second item of the second paragraph of the patent law.

2, starting from the background technology

When writing background technology, it generally introduces what algorithm is used in the prior art. What are the disadvantages of this algorithm, but the writing of a good background technology must be combined with the actual application field or engineering, where the application field or project must be specific. It is not simply written for use in the chemical engineering or engineering, electrical or engineering, mechanical or engineering fields. The background art is written in accordance with the above method, and the probability of the invention falling under Article 25, paragraph 1, item 2 of the Patent Law can be reduced.

3. Start with the example

When writing an embodiment, several embodiments may be listed in the specific embodiments, each of which is a different application field or engineering, which can solve different technical problems, and writing according to the above method can reduce the invention belonging to Article 25 of the Patent Law. Probability of paragraph 1 of paragraph 1.

4. Start with the writing of claims

Regardless of the above-mentioned problems, background art, and embodiments, the parameters of some of the algorithms in the claims are replaced with parameters or physical quantities that are closely related to a specific actual project or device, and sometimes applications are given in the background art. The technical field, but because the inventor pursued the scope of protection as much as possible, and then neglected the writing here, it is easy to cause the non-compliance with the provisions of Article 25, paragraph 1 and 2 of the Patent Law, as follows:

BACKGROUND OF THE INVENTION Support Vector Machines is a very effective machine learning technique for classification by Vapnik. Compared to other classification techniques, support vector machines have many unique advantages, such as nonlinear construction, high-dimensional classification functions using kernel functions, and good generalization capabilities. Classification technology refers to the division of objects into different entries or classes by creating rules. Rules are created primarily by building classifiers or classifiers using class elements known in the dataset. Once you have successfully built the classification function, you can apply it to a new data set with the unknown number of classes, and hope that the new classification is as accurate as possible. Support vector machines are used in many fields, such as feature recognition, text classification, face recognition, and file classification. Claims: A DNA encoding based support vector machine parameter prediction method, comprising:

Randomly generate DNA candidate populations with T chromosomes, and set control parameters: including chromosome length L, probability of transgenic and mutation operations, DNA population size T, maximum iteration number Gmax;

DNA decoding and support vector machine training: in the current population, each chromosome is decoded into corresponding adjustment parameters C and Gaussian kernel parameters, SVM training is performed with these parameter values, and the fitness value is calculated;

After calculating the fitness value, it is judged whether the local search condition is satisfied, and if it is satisfied, the local search is performed, otherwise the selection operation is performed;

When a selection operation is performed on the current population, a new chromosome population is generated, and a plurality of chromosomes are randomly generated in the new DNA population until the total number of chromosomes reaches T;

Performing a transgenic operation on the newly generated population; performing a mutation operation on the newly generated population;

Determine whether the termination condition is satisfied. If it is not satisfied, continue to perform DNA decoding and support vector machine training until the termination criterion is met, record the best chromosome, select the optimal adjustment parameter C and the Gaussian kernel parameter and stop running.

 The opinion given by the examiner is as follows: Even if the applicant modifies claim 1, the "support vector machine application field is characterized by feature recognition, text classification, face recognition or document classification", and the claims of the present application protect against DNA coding. Support vector machine parameter prediction method, the program defines only the description of the algorithm rules themselves, including the characteristics of the population, DNA and other mathematical concepts and terms in the algorithm, which does not reflect the feature recognition, text classification, face recognition. Or the specific data in the file classification structure does not reflect how the specific processing of the algorithm is related to the structure of the specific domain, nor does it define which constraint problem in the structure of the corresponding specific domain can be solved by the algorithm. The technical effect, therefore, according to the guidelines, the essence of the claim is still limited to an algorithm or mathematical calculation rules themselves, not to the object of patent protection. In summary, the author believes that the above cases must be written in conjunction with specific actual engineering or installation parameters or physical quantities when writing claims. For details, see patent CN201610788856.7 and CN201010209876.9.

Second, when replying to the review opinion

When answering the review opinions, we should start from the following three aspects: analysis from the target audience; analysis from technical means; analysis from technical effects.

From the analysis of the target: (for the answer)

It is to be noted that the present invention is specific to the object in the background art and the embodiment, and the object may be a specific actual project, or may be a specific device.

You can view the corresponding keywords in the text, and sometimes you can find the device corresponding to the keyword through keywords; for example, "space diversity", Baidu search shows that spatial diversity is also called antenna diversity, which is the most used in wireless communication. One of the diversity forms, in simple terms, is to use multiple receiving antennas to receive the signals and then combine them. Obviously, spatial diversity is studied for the electromagnetic scattering of the antenna.

It is possible to search for whether the algorithm has an application in other patents or projects. If there is an application, the invention is equivalent to an improvement to the existing algorithm, and this method can also be applied to a corresponding device or project.

Analyze from technical means:

Technical means are extracted from the entire scheme. The so-called technical means refers to the technical means that can be obtained by combining certain laws or parameters with natural laws.

Analysis from technical effects:

Technical effects are advantages compared to existing technical solutions, and the technical effects can be appropriately reasoned according to the technical solutions in the text.

Gowo Intellectual Property has a team of patent agents who combine theory and practice in a multi-technology field, which can meet the needs of all industries! Patent application to contact Gaowo intellectual property, professional, reliable!

Source: Beijing Gaowo Intellectual Property Public Number (ID: gaowoip-com)