Outlier detection method and related apparatus based on data collection and knowledge filtering

Abstract

The application belongs to the technical field of industrial production data outlier detection / abnormal point detection, and discloses an outlier detection method based on data collection and knowledge filtering and a related device; wherein the outlier detection method based on data collection and knowledge filtering comprises the following steps: collecting industrial production data; based on the collected industrial production data, using a knowledge filtering-based outlier detection model for classification detection to obtain an outlier detection result; wherein the knowledge filtering-based outlier detection model comprises an encoder, a decoder and a knowledge filter. The technical scheme provided by the application can solve the problem of low outlier detection precision caused by the lack of utilization of prior knowledge in the prior art, and can significantly improve the outlier detection precision compared with the prior art.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of industrial production data outlier detection / abnormal point detection, and particularly relates to an outlier detection method based on data collection and knowledge filtering and a related device. BACKGROUND

[0002] A large amount of production data is generated in industrial production processes, which can represent the quality of products to some extent. Further, in mass production processes, finding outliers in products based on production data can help enterprises improve product quality and effectively reduce a series of safety and economic problems caused by product failures.

[0003] At present, industrial production processes are extremely complex. Even if the signals are acquired in real time and through intensive collection, they still cannot fully represent product quality, and the outlier detection accuracy of existing detection schemes needs to be improved. In addition, experienced workers can filter outliers from products by combining their own knowledge with process documents, and have high detection accuracy, but the labor cost is high and the detection accuracy is limited by the experience level of workers.

[0004] In summary, due to the low outlier detection accuracy caused by the fact that the prior knowledge such as worker experience and process documents is not fully utilized in the prior art, there is an urgent need for a new outlier detection scheme. SUMMARY

[0005] The present application aims to provide an outlier detection method based on data collection and knowledge filtering and a related device to solve one or more of the above technical problems. The technical scheme provided by the present application can solve the problem of low outlier detection accuracy caused by the lack of utilization of prior knowledge in the prior art, and can significantly improve the outlier detection accuracy compared to the prior art.

[0006] To achieve the above-mentioned purpose, the present application adopts the following technical scheme:

[0007] In a first aspect of the present application, an outlier detection method based on data collection and knowledge filtering is provided, comprising the following steps:

[0008] Collecting industrial production data;

[0009] Based on the collected industrial production data, a knowledge filtering-based outlier detection model is used for classification detection to obtain an outlier detection result;

[0010] Among them,

[0011] The structure of the knowledge filtering-based outlier detection model is a neural network model, comprising:

[0012] An encoder is configured to input original industrial production data and perform dimension compression to obtain low-dimensional features.

[0013] A decoder is configured to map the low-dimensional features to the dimension of the original industrial production data to obtain reconstructed data.

[0014] A knowledge filter is configured to input the original industrial production data and the reconstructed data and extract features therefrom.

[0015] In the optimization of the knowledge filter-based outlier detection model, gradient backpropagation is performed according to the overall error, and the optimization target is to reduce the overall error.

[0016]

[0017] In the formula, E represents the overall error, E represents the reconstruction error, which is the mean square error between the input industrial production data and the reconstructed data, and E represents the filtering error, which is the error between the features of the input industrial production data and the reconstructed data extracted by the knowledge filter.

[0018] Further improvements of the present application are as follows.

[0019] The knowledge filter obtains the mathematical formula by inducing the industrial production process file and converting the text for judging product quality in the process file into the mathematical formula.

[0020] Further improvements of the present application are as follows.

[0021] The text for judging product quality includes one or both of the threshold type index and the mechanism index.

[0022] Further improvements of the present application are as follows.

[0023] The outlier detection result is the industrial production data whose overall error exceeds the preset threshold.

[0024] Alternatively, the outlier detection result is the industrial production data with the maximum overall error. N N is the product of the total number of collected industrial production data and the preset percentage.

[0025] In the second aspect of the present application, a data collection and knowledge filter-based outlier detection system is provided, which comprises:

[0026] A data collection module is configured to collect industrial production data.

[0027] ​​​​​The classification detection module is configured to perform classification detection based on the collected industrial production data by using a knowledge filter-based outlier detection model to obtain an outlier detection result.

[0028] Among them,

[0029] The knowledge filter-based outlier detection model is a neural network model, comprising:

[0030] The encoder is configured to input the original industrial production data and perform dimension compression to obtain low-dimensional features.

[0031] The decoder is configured to map the low-dimensional features to the dimension of the original industrial production data to obtain reconstructed data.

[0032] The knowledge filter is configured to input the original industrial production data and the reconstructed data and extract features therefrom.

[0033] When the knowledge filter-based outlier detection model is optimized, gradient backpropagation is performed according to the overall error, and the optimization target is to reduce the overall error; wherein the expression of the overall error is:

[0034] ;

[0035] In the formula, represents the overall error; represents the reconstruction error, which is the mean square error between the input industrial production data and the reconstructed data; represents the filtering error, which is the error between the features of the input industrial production data and the reconstructed data extracted by the knowledge filter.

[0036] Further improvements of the present application are as follows,

[0037] The knowledge filter obtains the mathematical formula by inducing the industrial production process file and converting the text for judging product quality in the process file into a mathematical formula; wherein the mathematical formula has one input and one output.

[0038] Further improvements of the present application are as follows,

[0039] The text for judging product quality includes one or both of threshold indicators and mechanism indicators.

[0040] Further improvements of the present application are as follows,

[0041] The outlier detection result is the industrial production data whose overall error exceeds the preset threshold;

[0042] Alternatively, the outlier detection result is the industrial production data with the largest overall error. N N ​a product of a total number of collected industrial production data and a preset percentage.

[0043] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for outlier detection based on data collection and knowledge filtering according to any one of the first aspect of the present application when executing the program.

[0044] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for outlier detection based on data collection and knowledge filtering according to any one of the first aspect of the present application.

[0045] Compared with the prior art, the present application has the following beneficial effects:

[0046] The outlier detection method disclosed by the present application is specifically a method for outlier detection based on data collection and knowledge filtering, which can solve the problem of low outlier detection precision caused by lack of utilization of prior knowledge in the prior art, and can significantly improve the outlier detection precision compared with the prior art. Specifically and by way of explanation, the outlier detection model based on knowledge filtering used in the present application is provided with a knowledge filter, which can fully utilize prior knowledge such as worker experience and process files, and these prior knowledge can effectively improve the detection effect and significantly improve the level of outlier detection; further, the method of the present application can accurately detect outliers in a batch of products, and can improve the overall consistency and quality of products, and has significant advantages and application potential in improving production efficiency, ensuring product quality, and promoting industrial upgrading, and has broad application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description; obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0048] Figure 1 is a flowchart of a method for outlier detection based on data collection and knowledge filtering in an embodiment of the present application;

[0049] Figure 2 is a schematic diagram of the principle of the embodiment shown in Figure 1

[0050] Figure 3 is a flowchart of a method for outlier detection based on data collection and knowledge filtering in a specific embodiment of the present application;​

[0051] Figure 4 is a schematic diagram of a data collection and knowledge filtering-based outlier detection system in an embodiment of the present application. DETAILED DESCRIPTION

[0052] For the purposes of the present application, the technical solutions and advantages will be more apparent, the technical solutions in the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application; obviously, the described embodiment technical solutions are a part of the embodiments of the present application, and not all the embodiments.

[0053] Based on the technical solutions disclosed in the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without making creative labor fall within the scope of protection of the present application. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0054] Please refer to Figure 1 and Figure 2 , the data collection and knowledge filtering-based outlier detection method disclosed in the embodiments of the present application comprises the following steps:

[0055] Step 1, collecting industrial production data;

[0056] Step 2, based on the industrial production data collected in step 1, using a knowledge filtering-based outlier detection model for classification detection to obtain an outlier detection result;

[0057] The structure of the knowledge filtering-based outlier detection model is a neural network model, comprising:

[0058] An encoder is used to input the original industrial production data and perform dimension compression to obtain low-dimensional features;

[0059] A decoder is used to map the low-dimensional features to the dimension of the original industrial production data to obtain reconstructed data;

[0060] A knowledge filter is used to input the original industrial production data and the reconstructed data and extract features respectively;

[0061] When the knowledge filtering-based outlier detection model is optimized, gradient backpropagation is performed according to the overall error; wherein the expression of the overall error is:

[0062] ;

[0063] wherein, represents the total error; represents the reconstruction error, the reconstruction error being the mean square error between the input industrial production data and the reconstructed data; represents the filtering error, the filtering error being the error between the input industrial production data extracted by the knowledge filter and the features of the reconstructed data.

[0064] The technical scheme provided by the embodiment of the application is used for solving the problem of low detection precision caused by lack of use of prior knowledge in the prior art, and can effectively improve the precision of outlier detection.

[0065] Please refer to Figure 3 In an embodiment of the application, the step of constructing the knowledge filtering-based outlier detection model can specifically include:

[0066] Step 2.1: Construct an outlier detection model without knowledge, which can be referred to as a basic model; wherein the basic model includes an input and an output, the input is multi-source, that is, it can accept multi-dimensional data; the basic model is a neural network model in structure, which can accept input data, compress the dimensions of the data through an encoder to obtain low-dimensional features of the data, and then map the low-dimensional features of the data to the original data dimensions through a decoder, at this time, the obtained data is referred to as reconstructed data or reconstruction result; interpretively, the mean square error between the reconstructed data and the input data is referred to as the reconstruction error, and the parameter optimization direction of the model is to make the reconstructed data and the input data as similar as possible; interpretively, the basis for the model to detect outliers in the basic model is that each sample data input into the model will produce a reconstructed data, and at the same time, a reconstruction error will be produced, and the reconstruction errors are arranged from large to small; there are two strategies to screen outlier samples, the first one is to set a threshold, when the reconstruction error of a certain sample exceeds the threshold, the sample is judged as an outlier; the second one is to set a percentage of α%, and the sample number of the largest reconstruction error × α% of the samples is classified as outliers;

[0067] Step 2.2: embed the knowledge into the outlier detection model of step 2.1 to construct a knowledge filtering-based outlier detection model; wherein the difference between the model constructed in step 2.2 and the outlier detection model without knowledge (i.e. the basic model) is that the model constructed in step 2.2 is the final model to be constructed in the technical scheme of the embodiment of the application, and the model of step 2.1 is a basis, which is used as a framework and prepares for embedding knowledge;

[0068] Step 2.3: Induction of process documents and conversion of text therein for judging product quality into mathematical formula, also known as knowledge-based feature filter, simply referred to as feature filter (knowledge filter), which has one input and one output, the dimension of the input data is consistent with the knowledge-based filter outlier detection model, and the output is determined by the prior knowledge; at this time, the reconstructed data obtained in step 2.1 is considered, since the reconstructed data and the input data are of the same dimension, it can be used as the input of the formula constructed in this step; at the same time, the target of data reconstruction is consistent with the input data, therefore, the reconstruction result and the input have high similarity and retain certain differences;

[0069] Step 2.4: In step 2.3, a knowledge-based feature filter is constructed, which extracts features from both the reconstruction result and the input data, which is equivalent to obtaining data features from another dimension, and the pattern of the features is determined by the knowledge from which the feature filter or the formula comes. In step 2.1, a basic model is constructed, which will perform gradient back propagation according to the reconstruction error, i.e. the mean square error between the reconstructed data and the input data, when optimizing; after extracting the features of the reconstruction result and the input data, the error between the two features is called filtering error, which is added to the loss function of the model in step 2.1, then the final optimization target of the model will be updated from "the parameter optimization direction of the model is to make the reconstructed data and the input data as similar as possible" in step 2.1 to "reduce the reconstruction error and the filtering error of the data at the same time", at this time, the knowledge-based filter outlier detection model is simply referred to as "knowledge filter model", and the error obtained is named "total error".

[0070] In one embodiment of the present application, the main process indicators can be summarized by analyzing the process documents, including threshold indicators and mechanism indicators, and converted into mathematical expressions, which have one input and one output, and the input data is a certain feature dimension of the production data. For example, the maximum voltage of a 18650 type battery in industrial production is 4.2V, which is converted and summarized into a mathematical expression to form a formula, when the input data contains voltage, if the voltage exceeds 4.2V, a signal 1 is sent, otherwise a signal 0 is sent.

[0071] In one embodiment of the present application, a data acquisition and knowledge filtering based outlier detection method implementation device is provided, which can include a data collector, a knowledge filtering based outlier detection model and a server; wherein the data collector is used to collect industrial production data, including real-time data and discrete data; the knowledge filtering based outlier detection model is deployed on the server (central server) to receive the data collected by the data collector, and then detect and classify the data to find the outlier samples.

[0072] Further example, the knowledge filtering based outlier detection model can be implemented as code and deployed on a computer, which can be referred to as a central server; wherein the source of data is a data collector, which will be transmitted to the central server through the network, and then the interface of the model is called by the server to input the data into the model; the model obtains the overall error of the data, and ranks according to the error, and detects and outputs a result in the manner specified in step 2.1, which is whether the sample in the data is an outlier sample.

[0073] The embodiment of the application proposes a knowledge filtering based outlier detection scheme, aiming to solve the defect of low detection precision caused by insufficient utilization of prior knowledge in outlier detection in the prior art, and can significantly improve the precision and efficiency of detection. Further, in a specific implementation system, three core functions of data collection, intelligent analysis and feedback processing can be integrated, which is suitable for multiple fields such as industrial production, quality control and signal processing, and realizes accurate outlier identification of various complex data; wherein the data collector is not only responsible for collecting real-time data and discrete data in industrial production, but also has data preprocessing capabilities such as data cleaning and format unification, to ensure that the data input into the detection model has high quality and strong accuracy, supports data collection from multiple sources (such as sensors, log files, databases, etc.), meets the data acquisition needs in different scenarios, and enhances the flexibility and applicability of the system; the core of the knowledge filtering based outlier detection model lies in the introduction of a target knowledge embedding model, which can integrate worker experience, process documents, historical cases and other prior knowledge to provide strong knowledge support for outlier detection. After receiving the data transmitted by the data collector, the model first performs feature extraction, and then intelligently classifies the data through a knowledge filtering algorithm to quickly and accurately identify outlier samples that deviate significantly from the normal mode. The model has self-learning and updating capabilities, and can continuously optimize the knowledge base and algorithm parameters according to newly collected feedback data and detection results, to realize continuous improvement of detection precision; the central server is responsible for carrying the outlier detection model, providing powerful computing power and storage space to ensure the efficiency and stability of large-scale data processing. The server can also display the detection results to the user in the form of intuitive charts and reports, and can automatically trigger an alarm mechanism according to a preset threshold to notify relevant personnel to take measures in a timely manner. In summary, the technical scheme of the application can effectively reduce false positives and false negatives, improve the accuracy of outlier detection, identify and process outliers in the production process in a timely manner, help improve the quality and consistency of the entire batch of products, and reduce the rate of defective products. The implementation of the scheme provides a powerful tool for the intelligentization and automation of industrial production, and promotes the transformation of manufacturing industry to intelligent manufacturing.

[0074] The following is an apparatus embodiment of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the apparatus embodiment, please refer to the method embodiment of the present application.

[0075] Please refer to Figure 4 In the embodiment of the present application, a data collection and knowledge filtering based outlier detection system is provided, comprising:

[0076] a data collection module, configured to collect industrial production data;

[0077] a classification detection module, configured to perform classification detection based on the collected industrial production data by using a knowledge filtering based outlier detection model to obtain an outlier detection result;

[0078] wherein,

[0079] The structure of the knowledge filtering based outlier detection model is a neural network model, comprising:

[0080] an encoder, configured to input the original industrial production data and perform dimension compression to obtain low-dimensional features;

[0081] a decoder, configured to map the low-dimensional features to the dimension of the original industrial production data to obtain reconstructed data;

[0082] a knowledge filter, configured to input the original industrial production data and the reconstructed data and extract features respectively;

[0083] When the knowledge filtering based outlier detection model is optimized, gradient backpropagation is performed according to the overall error, and the optimization goal is to reduce the overall error; wherein the expression of the overall error is:

[0084] ;

[0085] In the formula, represents the overall error; represents the reconstruction error, which is the mean square error between the input industrial production data and the reconstructed data; represents the filtering error, which is the error between the features of the input industrial production data and the reconstructed data extracted by the knowledge filter.

[0086] In an embodiment of the present application, a computer device is provided, which comprises a processor and a memory for storing a computer program comprising program instructions, the processor being configured to execute the program instructions stored in the computer storage medium. The processor can be a Central Processing Unit (CPU), and can also be other general-purpose processors, Digital Signal Processors (DSP), Application Specific Integrated Circuits (ASIC), Field-Programmable Gate Arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, and are particularly suitable for loading and executing one or more instructions in the computer storage medium to implement a corresponding method flow or a corresponding function. The processor in the embodiments of the present application can be used to execute the operations of the outlier detection method based on data collection and knowledge filtering.

[0087] In an embodiment of the present application, a storage medium is provided, specifically a computer readable storage medium (Memory), which is a memory device in a computer device, and is used to store programs and data. It can be understood that the computer readable storage medium herein can include an internal storage medium in the computer device, and of course can also include an extended storage medium supported by the computer device. The computer readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM (Random Access Memory) memory, or a non-volatile memory such as at least one disk memory. One or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the outlier detection method based on data collection and knowledge filtering in the above embodiments.

[0088] Those skilled in the art will appreciate that embodiments of the application can be devised for a method, a system, or a computer program product. Accordingly, the present application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage etc.) embodying computer readable program code.

[0089] The present application is described in reference to the flowchart and / or block diagram of the method, apparatus (system) and computer program product according to embodiments of the application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and a combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified in the flowchart and / or block diagram block or blocks.

[0090] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flowchart and / or block diagram block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified in the flowchart and / or block diagram block or blocks.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart and / or block diagram block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified in the flowchart and / or block diagram block or blocks.

[0092] Finally, it should be noted that the above-mentioned embodiments are merely intended for describing and illustrating, not limiting, the technical solutions of the present application. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. An outlier detection method based on data acquisition and knowledge filtering, characterized in that, Includes the following steps: Collect industrial production data; Based on the collected industrial production data, a knowledge-based outlier detection model is used for classification and detection to obtain outlier detection results. in, The outlier detection model based on knowledge filtering has a neural network structure, including: An encoder is used to input raw industrial production data and perform dimensionality compression to obtain low-dimensional features. A decoder is used to map the low-dimensional features to the original industrial production data dimensions to obtain reconstructed data. A knowledge filter is used to input raw industrial production data and reconstructed data and extract features respectively; The knowledge-based outlier detection model, during optimization, performs gradient backpropagation based on the overall error, with the optimization objective being to reduce the overall error; wherein the expression for the overall error is: ； In the formula, Indicates the total error; This represents the reconstruction error, which is the mean square error between the input industrial production data and the reconstructed data. This represents the filtering error, which is the error between the features extracted by the knowledge filter from the input industrial production data and the reconstructed data. The knowledge filter is obtained by summarizing industrial production process documents and converting the text in the process documents used to judge product quality into mathematical formulas; wherein, the mathematical formula has one input and one output.

2. The outlier detection method based on data acquisition and knowledge filtering according to claim 1, characterized in that, The text used to determine product quality includes one or both of threshold-based indicators and mechanistic indicators.

3. The outlier detection method based on data acquisition and knowledge filtering according to claim 1, characterized in that, The outlier detection result is industrial production data whose overall error exceeds a preset threshold; Alternatively, the outlier detection result is the one with the largest overall error. N Individual industrial production data; among them N This is the product of the total amount of industrial production data collected and a preset percentage.

4. An outlier detection system based on data acquisition and knowledge filtering, characterized in that, include: The data acquisition module is used to collect industrial production data; The classification and detection module is used to classify and detect outliers based on the collected industrial production data using a knowledge-based filtering outlier detection model, and obtain outlier detection results. in, The outlier detection model based on knowledge filtering has a neural network structure, including: An encoder is used to input raw industrial production data and perform dimensionality compression to obtain low-dimensional features. A decoder is used to map the low-dimensional features to the original industrial production data dimensions to obtain reconstructed data. A knowledge filter is used to input raw industrial production data and reconstructed data and extract features respectively; The knowledge-based outlier detection model, during optimization, performs gradient backpropagation based on the overall error, with the optimization objective being to reduce the overall error; wherein the expression for the overall error is: ； In the formula, Indicates the total error; This represents the reconstruction error, which is the mean square error between the input industrial production data and the reconstructed data. This represents the filtering error, which is the error between the features extracted by the knowledge filter from the input industrial production data and the reconstructed data. The knowledge filter is obtained by summarizing industrial production process documents and converting the text in the process documents used to judge product quality into mathematical formulas; wherein, the mathematical formula has one input and one output.

5. The outlier detection system based on data acquisition and knowledge filtering according to claim 4, characterized in that, The text used to determine product quality includes one or both of threshold-based indicators and mechanistic indicators.

6. The outlier detection system based on data acquisition and knowledge filtering according to claim 4, characterized in that, The outlier detection result is industrial production data whose overall error exceeds a preset threshold; Alternatively, the outlier detection result is the one with the largest overall error. N Individual industrial production data; among them N This is the product of the total amount of industrial production data collected and a preset percentage.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the outlier detection method based on data acquisition and knowledge filtering as described in any one of claims 1 to 3.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the outlier detection method based on data acquisition and knowledge filtering as described in any one of claims 1 to 3.

Images