JP7281054B2 - image forming device - Google Patents

Description

The present invention relates to an image forming apparatus.

When a problem such as a malfunction occurs in an image forming apparatus, it is necessary to quickly grasp the situation of the problem and identify the cause. Therefore, a technology has been proposed that acquires data for debugging (executes a debug log) only when there is a possibility that a problem will occur (see, for example, Patent Document 1).

JP 2011-210259 A

However, in order to execute the debug log, it is necessary for the developer to clarify detection conditions for detecting failures and incorporate them into the program in advance. However, when an image forming apparatus that is multi-functional and complicated performs an unexpected operation, it may not be possible to detect the defect. Furthermore, it is difficult to clarify all failures, including unexpected behavior, in advance, and there is room for improvement in grasping failures.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to provide an image forming apparatus capable of more comprehensively grasping defects.

According to the present invention, the image forming apparatus includes a first acquisition unit, a first symbolization unit, a learning unit, a second acquisition unit, a second symbolization unit, an inference unit, a determination unit, and a specification unit. an operation control unit; and a third acquisition unit. The first acquisition unit acquires first log information indicating an operation status of the image forming apparatus. The first symbolizing unit assigns a first symbol uniquely corresponding to the first log information. The learning unit inputs the first symbol to a learning model using a neural network, and causes the learning model to learn a first probability distribution indicating the appearance probability of the first log information. The second acquisition unit acquires second log information indicating the operation status of the image forming apparatus and having the same format as the first log information. The second symbolization unit assigns a second symbol uniquely corresponding to the second log information. The inference unit inputs the second symbol to the learning model, causes the learning model to output a second probability distribution indicating the appearance probability of the second log information, and based on the second probability distribution, the Infer the order of appearance of the second log information. The determining unit determines whether or not there is a possibility of malfunction of the image forming apparatus based on a comparison between the inferred order of appearance of the second log information and the actual order of appearance of the second log information. The identifying unit identifies an operation of the image forming apparatus that may be malfunctioning when the determining unit determines that the image forming apparatus may be malfunctioning. The operation control unit controls the operation of the image forming apparatus so as to reproduce the operation specified by the specifying unit. The third acquisition unit acquires third log information indicating an operation status of the image forming apparatus when the image forming apparatus reproduces the operation specified by the specifying unit.

ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to grasp|ascertain a malfunction more comprehensively.

1 is a schematic diagram showing an outline of an image forming apparatus according to an embodiment of the present invention; FIG. 1 is a block diagram showing the configuration of an image forming apparatus according to an embodiment; FIG. FIG. 3 is a diagram showing an example of software log information according to the embodiment and translation contents corresponding to the software log information; It is a figure which shows an example of the software log dictionary which concerns on this embodiment. It is a figure which shows an example of the corpus which concerns on this embodiment, and the binarization data of a corpus. 1 is a schematic diagram showing an example of a learning model including a recursive neural network according to this embodiment; FIG. It is a figure which shows an example of the defect estimation result table which concerns on this embodiment. 4 is a flowchart showing defect estimation processing according to the embodiment; It is a figure which shows an example of the log information comparison result table which concerns on this embodiment. It is a figure which shows an example of the log information comparison table which concerns on this embodiment. 6 is a flowchart showing log acquisition processing according to the embodiment; 6 is a flowchart showing a part of detailed log acquisition processing according to the embodiment; 8 is a flowchart showing another part of detailed log acquisition processing according to the present embodiment; 9 is a flowchart showing still another part of detailed log acquisition processing according to the present embodiment;

Embodiments of the present invention will be described below with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals, and description thereof will not be repeated.

First, an outline of an image forming apparatus 100 according to the present embodiment will be described with reference to FIG. FIG. 1 is a schematic diagram showing an outline of an image forming apparatus 100. As shown in FIG. The image forming apparatus 100 forms an image on the sheet S. FIG. The image forming apparatus 100 is, for example, a printer, a copier, or a multifunction machine. Image forming apparatus 100 may have a facsimile function. In this embodiment, the image forming apparatus 100 is electrophotographic.

Image forming apparatus 100 includes image forming section 110 , acquisition section 120 , communication section 130 , control section 140 , input/output section 150 , feeding section 112 , and conveying section 114 . Image forming unit 110 , acquisition unit 120 , communication unit 130 and control unit 140 are installed inside the housing of image forming apparatus 100 .

Acquisition unit 120 acquires log information indicating the operation status of image forming apparatus 100 . Specifically, the acquiring unit 120 acquires software log information indicating the operation status of the image forming apparatus 100 while the image forming apparatus 100 forms an image on the sheet S, for example. Acquisition unit 120 also acquires, for example, software log information indicating the operation status of image forming apparatus 100 during warm-up operation of image forming apparatus 100 . The warm-up operation includes, for example, an operation of raising the temperature of the heating roller 110j to a temperature at which the fixing operation can be performed, and a roller forming the developing unit 110d before the feeding roller 112b starts feeding the sheet S. and at least one of the action of rotating the etc. for a certain period of time.

Here, the “software log information” is information in which the behavior of the image forming apparatus 100 is recorded in chronological order along with date and time. For example, the software log information indicates the state of the image forming apparatus 100, the control result of each function, the state of an error or failure, or the state of communication with other electronic devices. Acquisition unit 120 includes first acquisition unit 122 , second acquisition unit 124 , third acquisition unit 126 , and fourth acquisition unit 128 .

First acquisition unit 122 acquires first log information indicating the operating status of image forming apparatus 100 . Here, the "first log information" is software log information used for learning the first probability distribution by the same technique as the learning model for natural language processing. In this embodiment, the first log information includes, for example, approximately 3000 to 4000 pieces of software log information. The first probability distribution indicates the appearance probability of each piece of software log information included in the first log information.

A second acquisition unit 124 acquires second log information indicating the operation status of the image forming apparatus 100 . Here, the "second log information" is software log information that is input to the trained learning model and used to output the second probability distribution. In this embodiment, the second log information includes, for example, approximately 1000 to 1500 pieces of software log information. A second probability distribution indicates the appearance probability of each piece of software log information included in the second log information. The second log information and the first log information have the same format, but different contents.

Third acquisition unit 126 acquires third log information indicating the operation status of image forming apparatus 100 . Here, the "third log information" is software log information used for comparing the actual appearance order of the second log information and the actual appearance order of the third log information. The first log information and the third log information have the same format, but different contents. Also, the second log information and the third log information have the same format, but may differ in content or may have the same content.

A fourth acquisition unit 128 acquires fourth log information indicating the operation status of the image forming apparatus 100 . Here, the "fourth log information" is software log information for a user or a service person to analyze an operation of the image forming apparatus 100 that may have a problem. The first log information and the fourth log information have the same format, but different contents. Further, the second log information, the third log information, and the fourth log information each have the same format, but may differ in content or may have the same content.

The communication unit 130 can communicate with an electronic device equipped with a communication device using the same communication method (protocol). Specifically, the communication unit 130 communicates with other electronic devices via a network such as a LAN (Local Area Network). The communication unit 130 is, for example, a communication module (communication device) such as a LAN board.

Control unit 140 controls the operation of each component of image forming apparatus 100 by executing a control program. Furthermore, the control unit 140 estimates the possibility of malfunction of the image forming apparatus 100 using a learning model including a neural network. Specifically, the control unit 140 inputs a symbol indicating software log information to a recurrent neural network (hereinafter sometimes referred to as “RNN”), and generates a learning model for natural language processing. Presence or absence of the possibility of malfunction of the image forming apparatus 100 is estimated using a method similar to that of . That is, the neural network includes a recurrent neural network. Therefore, the log information recorded in chronological order can be treated as input data to the learning model. For example, in the RNN, if peculiar software log information is detected, it is estimated that there is a possibility of malfunction. An RNN has an input layer, a hidden layer, and an output layer. The number of hidden layers may be one, or two or more.

The input/output unit 150 has a display unit 152 , a reception unit 154 and an audio output unit 156 .

Display unit 152 includes a display and a touch sensor. The display unit 152 is, for example, a touch panel including a liquid crystal display. The display is not limited to the liquid crystal display, and may be an organic EL display (Organic Electro Luminescence Display). A touch sensor detects a touch by an object to be detected. A touch sensor outputs a detection signal indicating a position where a touch by an object to be detected is detected. The object to be detected is, for example, a finger of the user. The touch sensor is, for example, a resistive touch sensor.

The display unit 152 notifies the user of various information by displaying various information on the display. Specifically, when control unit 140 determines that image forming apparatus 100 may be malfunctioning, display unit 152 displays information indicating that image forming apparatus 100 may be malfunctioning.

The reception unit 154 includes, for example, a numeric keypad, a start key, and a cancel key. At least one of accepting unit 154 and the touch sensor accepts an instruction from the user.

The voice output unit 156 notifies various information by voice. Specifically, when control unit 140 determines that image forming apparatus 100 may be malfunctioning, audio output unit 156 notifies that image forming apparatus 100 may be malfunctioning.

The feeding section 112 includes a cassette 112a and a feeding roller 112b. The cassette 112a accommodates the sheets S. As shown in FIG. The feeding roller 112b feeds the sheets S accommodated in the cassette 112a one by one as needed. When the cassette 112a accommodates a plurality of sheets S, the feed roller 112b feeds the uppermost sheet S one by one among the plurality of sheets S accommodated in the cassette 112a. Here, the feeding section 112 includes a plurality of cassettes 112a, and a feeding roller 112b is installed for each of the plurality of cassettes 112a. The sheet S is, for example, plain paper, recycled paper, thin paper, thick paper, coated paper, or an OHP (Overhead Projector) sheet.

The conveying portion 114 conveys the sheet S fed by the feeding portion 112 to the image forming portion 110 . Specifically, the conveying unit 114 conveys the sheets S one by one from the cassette 112 a toward the image forming unit 110 . After the image forming section 110 forms an image on the sheet S, the conveying section 114 conveys the sheet S from the image forming section 110 and discharges the sheet S to the outside of the image forming apparatus 100 .

The transport section 114 includes a plurality of transport rollers 114a. The transport roller 114a transports the sheet S. As shown in FIG. A transport path for the sheet S is formed by a plurality of transport rollers 114a.

Conveying roller 114a includes a rotating roller. The rotating roller rotates around a rotation axis. Typically, transport roller 114a includes a pair of rotating rollers. A pair of rotating rollers rotate around the rotating shaft while facing each other. In one example, one rotating roller of a pair of rotating rollers rotates according to the power of a motor, and the other rotating roller follows and rotates. The sheet S enters between a pair of rotating rollers and is pushed out from the rotating rollers by being biased by the rotating rollers.

The transport rollers 114a include registration rollers 114r. The registration roller 114 r adjusts the timing of conveying the sheet S to the image forming section 110 . The registration roller 114r temporarily stops conveying the sheet S, and conveys the sheet S to the image forming unit 110 in accordance with a predetermined timing of the image forming unit 110. FIG.

The toner containers Ca to Cd are attached to the image forming apparatus 100 . Each of toner containers Ca to Cd is detachable from image forming apparatus 100 . Each of the toner containers Ca to Cd contains toner of a different color. The toner in toner containers Ca to Cd is supplied to image forming section 110 . The image forming section 110 forms an image using toner supplied from the toner containers Ca to Cd.

For example, the toner container Ca accommodates yellow toner and supplies the yellow toner to the image forming unit 110 . The toner container Cb contains magenta toner and supplies the image forming unit 110 with the magenta toner. The toner container Cc accommodates cyan toner and supplies the image forming unit 110 with the cyan toner. The toner container Cd accommodates black toner and supplies the black toner to the image forming unit 110 . Toner is an example of a "developer."

The image forming unit 110 forms an image on the sheet S based on the image data using the toner contained in the toner containers Ca to Cd. Here, the image forming unit 110 includes an exposure unit 110a, a photosensitive drum 110b, a charging unit 110c, a developing unit 110d, a primary transfer roller 110e, a cleaning unit 110f, an intermediate transfer belt 110g, a secondary transfer roller 110h, and a fixing unit. Includes portion 110i. The photoreceptor drum 110b is an example of an "image carrier".

The intermediate transfer belt 110g is rotated by a rotating roller that rotates according to power of a motor. A motor is attached to the developing unit 110d. The toner in the developing section 110d is agitated as the motor rotates.

The photosensitive drum 110b, charging section 110c, developing section 110d, primary transfer roller 110e, and cleaning section 110f are provided corresponding to each of the toner containers Ca to Cd. A plurality of photoreceptor drums 110b are in contact with the outer surface of the intermediate transfer belt 110g and arranged along the rotation direction of the intermediate transfer belt 110g. A plurality of primary transfer rollers 110e are provided corresponding to a plurality of photosensitive drums 110b. A plurality of primary transfer rollers 110e face a plurality of photosensitive drums 110b via an intermediate transfer belt 110g.

The charging section 110c charges the peripheral surface of the photoreceptor drum 110b. The exposure unit 110a exposes each of the charged photosensitive drums 110b based on the image data. As a result, an electrostatic latent image is formed on the peripheral surface of the photosensitive drum 110b. The developing unit 110d develops the electrostatic latent image by supplying the toner to the exposed image bearing member to adhere the toner to the electrostatic latent image, thereby forming a toner image on the peripheral surface of the photosensitive drum 110b. . Accordingly, the photosensitive drum 110b carries a toner image. A primary transfer roller 110e transfers the toner image formed on the photosensitive drum 110b to the outer surface of the intermediate transfer belt 110g. The cleaning portion 110f removes toner remaining on the peripheral surface of the photoreceptor drum 110b.

The photosensitive drum 110b corresponding to the toner container Ca forms a yellow toner image based on the electrostatic latent image, and the photosensitive drum 110b corresponding to the toner container Cb forms a magenta toner image based on the electrostatic latent image. to form The photosensitive drum 110b corresponding to the toner container Cc forms a cyan toner image based on the electrostatic latent image, and the photosensitive drum 110b corresponding to the toner container Cd forms a black toner image based on the electrostatic latent image. to form

Toner images of a plurality of colors are superimposed and transferred from the photosensitive drum 110b onto the outer surface of the intermediate transfer belt 110g to form an image. As a result, the intermediate transfer belt 110g carries the image. The secondary transfer roller 110h transfers the image formed on the outer surface of the intermediate transfer belt 110g onto the sheet S.

The fixing unit 110i fixes the image to the sheet S by heating and pressing the sheet S onto which the image has been transferred. The fixing section 110i includes a heating roller 110j and a pressure roller 110k. Heat roller 110j and pressure roller 110k are positioned opposite each other to form a fusing nip. The sheet S that has passed between the intermediate transfer belt 110g and the secondary transfer roller 110h is pressed while being heated at a predetermined fixing temperature by passing through the fixing nip. As a result, the image is fixed on the sheet S. Conveyance unit 114 discharges sheet S on which the image is fixed to the outside of image forming apparatus 100 .

Next, the configuration of the image forming apparatus 100, particularly the configuration of the control unit 140, will be described in detail with reference to FIGS. 1 and 2. FIG. FIG. 2 is a block diagram showing the configuration of the image forming apparatus 100. As shown in FIG.

As shown in FIG. 2 , the controller 140 includes a processor 142 and a memory 144 . The processing unit 142 is, for example, a processor. The processor is, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a GPGPU (General-purpose computing on graphics processing units). Processing unit 142 controls the operation of each component of image forming apparatus 100 by executing a control program stored in storage unit 144 .

The processing unit 142 includes a symbolizing unit 1421 , a learning unit 1422 , an inference unit 1423 , a determination unit 1424 and an operation control unit 1425 . In this embodiment, the processing unit 142 executes the control program stored in the storage unit 144 so that the control program includes a symbolization unit 1421, a learning unit 1422, an inference unit 1423, a determination unit 1424, and an operation A function with the control unit 1425 is realized.

The symbolizing unit 1421 assigns a symbol uniquely corresponding to the information of one sentence of the software log dictionary, which will be described later with reference to FIG. Specifically, the symbolization unit 1421 assigns an ID number to one-sentence information that may be acquired as software log information. In this embodiment, an ID number is assigned to one sentence of information that may be acquired as the first log information to the fourth log information among the plurality of sentences of information in the software log dictionary. However, symbols may be given to all sentences of information in the software log dictionary. The symbolization unit 1421 creates a software log dictionary that collects all software log information to which ID numbers have been assigned. Note that the ID number is an example of the "identification number".

The encoding section 1421 includes a first encoding section 1421a, a second encoding section 1421b, a third encoding section 1421c, and a fourth encoding section 1421d. The first symbolizing unit 1421a uses the ID number assigned to the software log dictionary to assign a first symbol uniquely corresponding to the first log information. The second symbolization unit 1421b uses the ID number assigned to the software log dictionary to assign a second symbol uniquely corresponding to the second log information. The third symbolization unit 1421c uses the ID number assigned to the software log dictionary to assign a third symbol uniquely corresponding to the third log information. The fourth symbolization unit 1421d uses the ID number assigned to the software log dictionary to assign a fourth symbol uniquely corresponding to the fourth log information.

The learning unit 1422 inputs a first symbol to a learning model using a neural network, and causes the learning model to learn a first probability distribution indicating the appearance probability of first log information corresponding to the first symbol. Specifically, the learning unit 1422 inputs the first symbol to the learning model using the RNN, and causes the learning model to learn the first probability distribution.

In this embodiment, the learning unit 1422 sequentially inputs all the first symbols to the learning model including the RNN, and causes the learning model to learn the first probability distribution indicating the appearance probability of the first log information. Specifically, the learning unit 1422 uses a so-called “one-hot vector” to sequentially input all the first symbols to the learning model, and causes the learning model to learn the appearance probabilities of all the first symbols. At that time, the learning unit 1422 calculates an error between the order of appearance of the first symbols indicated by the output result output from the learning model and the order of appearance of the first symbols, which is teacher data. The learning unit 1422 updates so-called “weight” and “bias”, which are parameters of the learning model, based on the calculated error. In this way, the learning unit 1422 learns the appearance probabilities of all first symbols.

The inference unit 1423 inputs the second symbol to the learning model including the RNN and causes the learning model to output a second probability distribution indicating the appearance probability of the second log information. Then, the inference unit 1423 infers the appearance order of the second log information based on the second probability distribution output to the learning model. Specifically, the inference unit 1423 infers a second symbol appearing after a specific second symbol based on the second probability distribution output by the learning model. That is, the inference unit 1423 infers another second symbol appearing next to one second symbol for each of all the second symbols. The inference unit 1423 infers the appearance order of the second log information and derives the inference appearance order of the second log information. That is, the inference unit 1423 infers the inference appearance order of the second log information.

The determining unit 1424 determines whether or not there is a possibility of malfunction of the image forming apparatus 100 based on a comparison between the inferred order of appearance of the second log information and the actual order of appearance of the second log information.

Operation control unit 1425 controls each component of image forming apparatus 100 by executing a control program stored in storage unit 144 . That is, the operation control unit 1425 controls operations of the image forming apparatus 100 . The operation control unit 1425 includes arithmetic elements. Arithmetic elements include processors. A processor includes, for example, a central processing unit (CPU). The processor may include an Application Specific Integrated Circuit (ASIC).

The storage unit 144 stores various data and control programs. The storage unit 144 is configured by, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), and/or a Solid State Drive (SSD). Storage unit 144 may include an external memory. The external memory is removable media. The storage unit 144 may include, for example, a USB (Universal Serial Bus) memory and/or an SD (Secure Digital) card as an external memory.

Furthermore, the storage unit 144 non-temporarily records the control program in a computer-readable recording medium. Computer-readable recording media on which the control program is non-temporarily recorded include ROM (Read Only Memory), RAM (Random Access Memory), CD-ROM, magnetic tape, magnetic disk, or optical data storage device.

Next, a defect estimation method by the image forming apparatus 100 of the present embodiment will be described with reference to FIGS. 1 to 7. FIG. FIG. 3 is a schematic diagram for explaining the software log information 310 used in the defect estimation method by the image forming apparatus 100. As shown in FIG. FIG. 3 is a diagram showing an example of software log information 310 and translation content 320 of the software log information 310. As shown in FIG.

As described with reference to FIG. 2 , first acquisition unit 122 acquires software log information 310 as first log information indicating the operating status of image forming apparatus 100 . The first acquisition unit 122 stores the acquired first log information in a RAM built in the CPU or a non-temporary recording medium. The second acquisition unit 124 to the fourth acquisition unit 128 similarly acquire the software log information 310 as the second log information to the fourth log information, and store the acquired second log information to the fourth log information in the RAM, or Store in a non-temporary recording medium.

For the sake of convenience, FIG. 3 shows the translation content 320 of the software log information 310 corresponding to the software log information 310 . The translation content 320 of the software log information 310 indicates the content of the software log information 310 in an easy-to-understand manner for the user.

FIG. 4 is a diagram showing an example of the software log dictionary 400 according to this embodiment. As shown in FIG. 4, software log dictionary 400 includes ID number 410 and dictionary content 420 .

The ID number 410 is, for example, a number in ascending order. An ID number 410 is assigned to uniquely correspond to each dictionary content 420 .

The dictionary content 420 includes all single-sentence information that appears in the software log information 310 . The encoding unit 1421 assigns an ID number 410 to each piece of software log information constituting the dictionary content 420 (each piece of software log information to which an ID number 410 is assigned is hereinafter sometimes referred to as a "word").

FIG. 5 is a diagram showing an example of a corpus 510 and binarized data 520 of the corpus 510 according to this embodiment. Here, in a learning model for natural language processing, the term "corpus" refers to an example of a sequence of phrases observed in a certain language or in a specific field of that language. Moreover, a corpus is a database containing a large number of phrases searchable by a computer. The "corpus" of this embodiment is also stored in a database, like the corpus in the learning model for natural language processing.

FIG. 6 is a schematic diagram showing an example of a learning model including a recurrent neural network 600 according to this embodiment. Recurrent neural network 600 includes input layer 612 , hidden layer 613 , output layer 614 and softmax 615 . In this embodiment, input layer 612 receives, for example, a first symbol. Specifically, the input layer 612 is sequentially input with all first symbols represented using so-called "one-hot vectors". Note that the first symbol may be input to the input layer 612 as it is, without using the "one-hot vector" to represent the first symbol.

Hidden layer 613 is responsible for propagating values from input layer 612 to output layer 614 . Specifically, the hidden layer 613 propagates values from the input layer 612 to the output layer 614 by a function including "weight" and "bias". The output layer 614 outputs the probability that the word input to the input layer 612 will appear next. Softmax 615 indicates the softmax function. The softmax 615 converts the output values into values between 0 and 1, and calculates each output value so that the sum of the output values becomes "1".

Further, FIG. 6 shows RNN 610 at "time T," RNN 620 at "time T+1," RNN 630 at "time T+2," RNN 640 at "time T+3," and RNN 650 at "time T+4."

As shown in FIG. 6, the output value of the output layer 614 of the RNN 610 at "time T" is input to the hidden layer 613 of the RNN 620 at "time T+1". As a result, the strength of the relationship between the software log information located before and after is sequentially handed over to the software log information located behind.

FIG. 7 is an example of a defect estimation result table 700 according to this embodiment. The defect estimation result table 700 includes an index 701 , an inference result ID 702 , a second log ID 703 , and good/not good 704 . The index 701 indicates the order of each piece of software log information that constitutes the inference appearance order of the second log information. The inference result ID 702 indicates the second symbol of each of the plurality of pieces of software log information forming the inference appearance order. The second log ID 703 indicates the second symbol of each of the multiple pieces of software log information constituting the actual appearance order of the second log information. A good/not good 704 indicates the result of comparing the inference result ID 702 and the second log ID 703 . If the good/non-good 704 is “good”, it indicates that the possibility that the operation of the image forming apparatus 100 is defective is low. If the good/not good 704 is “not good”, it indicates that there is a high possibility that the image forming apparatus 100 has a malfunction.

As shown in FIG. 7, among the indexes 701 of "679" to "687", the second log IDs 703 with the indexes 701 of "683", "684" and "686" are judged to be "not good". . That is, the second log information with the index 701 of "683", "684", and "686" indicates that the operation of the image forming apparatus 100 is likely to have a problem.

Next, defect estimation processing of the image forming apparatus 100 of the present embodiment will be described with reference to FIGS. 1 to 8. FIG. FIG. 8 is a flowchart showing defect estimation processing of the image forming apparatus 100 of this embodiment. Here, the “failure estimation process” refers to a process of estimating a failure caused by software or the like of the image forming apparatus 100 . The defect estimation process is executed in steps S2 to S30.

Step S2: The operation control unit 1425 receives an instruction to prepare for learning or execute estimation from the user via the receiving unit 154 . When the operation control unit 1425 determines that an instruction to prepare for learning has been received from the user ("preparing for learning" in step S2), the process proceeds to step S4. When the operation control unit 1425 determines that an instruction to execute estimation has been received from the user ("execute estimation" in step S2), the process proceeds to step S20.

Step S4: The encoding unit 1421 generates the software log dictionary 400 that collects all software log information to which ID numbers have been assigned. The process proceeds to step S6.

Step S6: The first acquisition unit 122 acquires first log information. The symbolizing unit 1421 uses the ID number of the software log dictionary 400 for each piece of software log information of the first log information. The process proceeds to step S8.

Step S8: The operation control unit 1425 separates the corpus into training, cross-validation, and testing corpora. The process proceeds to step S10.

Step S10: The learning unit 1422 inputs the ID number of the first log information to the RNN. The process proceeds to step S12.

Step S12: The learning unit 1422 calculates the error. Specifically, the RNN calculates the error. The process proceeds to step S14.

Step S14: The learning unit 1422 determines whether the error is greater than the threshold. Specifically, the RNN calculates the error. If the learning unit 1422 determines that the error is greater than the threshold (Yes in step S14), the process proceeds to step S16. If the learning unit 1422 determines that the error is not greater than the threshold (No in step S14), the process proceeds to step S18.

Step S16: The learning unit 1422 updates the weight value and the bias value. Specifically, the RNN updates the weight and bias values. The process returns to step S10.

Step S18: The learning unit 1422 stores the weight value and the bias value. Specifically, the RNN stores weight and bias values. As a result, a learned RNN is generated. Processing ends.

Step S20: The second acquisition unit 124 acquires second log information. The encoding unit 1421 uses the ID number of the software log dictionary 400 for each piece of software log information of the second log information. The process proceeds to step S22.

Step S22: The inference unit 1423 inputs the second symbol of the second log information to the learned RNN and performs inference. Specifically, the inference unit 1423 inputs the second symbol of the second log information to the learned RNN. The inference unit 1423 causes the RNN to output the second probability distribution. The inference unit 1423 performs inference based on the second probability distribution output to the RNN, and derives the inference appearance order of the second log information as an inference result. The process proceeds to step S24.

Step S24: The determination unit 1424 determines whether or not the inference order of appearance of the second log information indicated by the inference result matches the actual appearance order of the second log information. The process proceeds to step S26.

Step S26: If the determination unit 1424 determines that the inference result and the output log match (Yes in step S26), the process ends. If the determination unit 1424 determines that the inference result and the output log do not match (No in step S26), the process proceeds to step S28.

Step S28: The audio output unit 156 notifies the user by an alarm sound or the like. Proceed to step S30.

Step S30: The image forming apparatus 100 executes log acquisition processing. Processing ends.

Next, log acquisition processing will be described with reference to FIGS. 2, 7, 9, and 10. FIG. When the determining unit 1424 determines that there is a possibility of malfunction of the image forming apparatus 100, the identifying unit 1426 identifies the operation of the image forming apparatus 100 that may be malfunctioning. Specifically, for example, the identifying unit 1426 refers to the failure estimation result table 700 to find a comparison result that is “non-favorable”. Next, the identifying unit 1426 can identify the operation of the image forming apparatus 100, which appears before the software log information of the second log ID 703 corresponding to the comparison result of "not good" among the second log IDs 703. and the software log information closest to the software log information of the second log ID 703 corresponding to the comparison result of "not good". The operation of the image forming apparatus 100 specified by the specifying unit 1426 is, for example, a warm-up operation or an operation in which the image forming unit 110 forms an image on the sheet S (hereinafter sometimes referred to as a “printing operation”). be.

The operation control unit 1425 controls operations of the image forming apparatus 100 so as to reproduce the operations specified by the specifying unit 1426 . For example, if the operation specified by the specifying unit 1426 is a printing operation, the operation control unit 1425 controls the operation of the image forming apparatus 100 so as to reproduce the printing operation. Further, for example, when the motion specified by the specifying unit 1426 is a warm-up motion, the motion control unit 1425 controls the motion of the image forming apparatus 100 so as to reproduce the warm-up motion.

Third acquisition unit 126 acquires third log information indicating the operation status of image forming apparatus 100 when image forming apparatus 100 reproduces the operation specified by specifying unit 1426 . Therefore, by comparing the actual appearance order of the second log information and the actual appearance order of the third log information, it is possible to determine whether the operation of the image forming apparatus 100 that may be defective and the operation of the image forming apparatus 100 in another state are detected. action can be compared. As a result, the serviceman can easily identify the cause of the malfunction of the image forming apparatus 100 .

FIG. 9 shows an example of a log information comparison result table 800 according to this embodiment. The log information comparison result table 800 includes an index 801 , a second log ID 802 , a third log ID 803 and match/mismatch 804 . The index 801 indicates the order of each piece of software log information that constitutes the actual appearance order of the second log information. The second log ID 802 indicates the second symbol of each of a plurality of pieces of software log information constituting the actual appearance order of the second log information. The third log ID 803 indicates the third symbol of each of the plurality of software log information constituting the actual appearance order of the third log information. A match/mismatch 804 indicates the result of comparing the second log information and the third log information. Specifically, match/mismatch 804 indicates the result of comparing the second symbol of second log ID 802 and the third symbol of third log ID 803 . If the match/mismatch 804 is "match", the second symbol of the second log ID 802 and the third symbol of the third log ID 803 match. If the match/mismatch 804 is "mismatch", the second symbol of the second log ID 802 and the third symbol of the third log ID 803 do not match.

As shown in FIG. 9, among the indices 801 of "679" to "687", the comparison results of the indices 801 of "684" and "686" are determined as "mismatch".

The determination unit 1424 determines whether or not the actual appearance order of the second log information and the actual appearance order of the third log information are similar. Specifically, the determining unit 1424 compares the actual appearance order of the second symbols of the second log information with the actual appearance order of the third symbols of the third log information, and determines the actual appearance order of the second log information. It is determined whether or not the order approximates the actual appearance order of the third log information. More specifically, for example, the determination unit 1424 refers to the log information comparison result table 800, and if the ratio of “mismatch” in the match/mismatch 804 is equal to or less than a predetermined value, the second log information It is determined that the actual appearance order and the actual appearance order of the third log information are close to each other. Further, for example, the determination unit 1424 refers to the log information comparison result table 800, and if the number of mismatches is equal to or less than a predetermined number, the actual appearance order of the second log information and the actual appearance order of the third log information is determined to be approximate. Therefore, the determination unit 1424 can easily determine the actual appearance order of the second log information and the third log information, rather than comparing the actual appearance order of the second log information and the actual appearance order of the third log information in the state of the corpus 510 . It can be determined whether or not the order of appearance of the log information is similar to the actual order of appearance.

When the determination unit 1424 determines that the actual appearance order of the second log information and the actual appearance order of the third log information are similar, the operation control unit 1425 controls the operation of the image forming apparatus 100 specified by the specifying unit 1426. The operation of the image forming apparatus 100 is controlled so as to reproduce the operation specified by the specifying unit 1426 by changing the conditions. Then, the fourth obtaining unit 128 obtains fourth log information indicating the operation status of the image forming apparatus 100 when reproducing the operation specified by the specifying unit 1426 by changing the operating conditions. Therefore, by comparing the actual appearance order of the fourth log information and the actual appearance order of the second log information with different operating conditions, the service person can easily identify the cause of the malfunction of the image forming apparatus 100 .

Reproduction of the motion specified by the specifying unit 1426 will be specifically described. When the operation specified by the specifying unit 1426 is the printing operation, the operation control unit 1425 changes the operating conditions of the printing operation and causes the image forming apparatus 100 to reproduce the printing operation. Specifically, if the size of the sheet S is A4 in the operating conditions of the printing operation when there is a possibility of a problem, the operation control unit 1425 changes the size of the sheet S to B5, and changes the size of the sheet S to B5. 100 to reproduce the printing operation. In addition, the operation control unit 1425 may change the number of sheets S to be printed in the operation conditions. Therefore, the service person can determine the image forming apparatus based on the difference between the operating conditions of the second log information and the operating conditions of the fourth log information, the actual order of appearance of the second log information, and the actual order of appearance of the fourth log information. 100 failure causes can be considered. As a result, the service person can more easily identify the cause of the malfunction of the image forming apparatus 100 .

Here, when the operation specified by the specifying unit 1426 is the printing operation, the operation control unit 1425 controls the transport unit 114 so as not to transport the sheet S from the cassette 112a toward the image forming unit 110, and the exposure unit 110a. forms an electrostatic latent image on the photosensitive drum 110b, thereby causing the image forming apparatus 100 to reproduce the printing operation. Therefore, the image forming section 110 can reproduce the printing operation without using consumable materials such as toner and the sheet S. FIG. As a result, the cost for identifying the cause of the malfunction of the image forming apparatus 100 can be reduced.

Further, when the operation specified by the specifying unit 1426 is the printing operation, the operation control unit 1425 controls the transport unit 114 so as not to transport the sheet S from the cassette 112a toward the image forming unit 110, and the exposure unit 110a The image forming unit 110 is controlled such that an electrostatic latent image is formed on the photoreceptor drum 110b and the developing unit 110d supplies toner to the photoreceptor drum 110b, thereby causing the image forming apparatus 100 to reproduce the printing operation. Therefore, the reproducibility of the printing operation is higher than that of reproducing the printing operation without supplying toner to the photosensitive drum 110b. As a result, the fourth log information that makes it easier to identify the cause of the malfunction of the image forming apparatus 100 can be obtained.

Furthermore, when the motion specified by the specifying unit 1426 is a warm-up motion, the motion control unit 1425 changes, for example, the pattern of the warm-up motion in the operating conditions. Specifically, the operation control unit 1425 controls the operation of the image forming apparatus 100 so that the warm-up operation includes the calibration operation, thereby changing the operating conditions and reproducing the warm-up operation in the image forming apparatus 100. Let Further, the operation control unit 1425 controls the operation of the image forming apparatus 100 so that the warm-up operation includes the adjusting operation of the charging unit 110c, thereby changing the operating conditions and reproducing the warm-up operation in the image forming apparatus 100. Let

Until the fourth acquisition unit 128 acquires a predetermined number of patterns of fourth log information, the operation control unit 1425 changes the operation conditions and causes the image forming apparatus 100 to reproduce the operation. The predetermined number is 3, for example. When the predetermined number is 3, the operation control section 1425 changes the operation condition 3 times. Then, the fourth acquisition unit 128 acquires three patterns of fourth log information. The predetermined number can be arbitrarily set by the user.

Next, a log information comparison table will be described with reference to FIG. FIG. 10 is a diagram showing a log information comparison table 900. As shown in FIG. The log information comparison table 900 includes an index 901 , an inference result ID 902 , a second log ID 903 , a third log ID 904 and a fourth log ID 905 . The index 901 indicates the order of each piece of software log information that constitutes the actual appearance order of the second log information. The inference result ID 902 indicates the second symbol of each of the plurality of pieces of software log information forming the inference appearance order of the second log information. The second log ID 903 indicates the second symbol of each of the multiple pieces of software log information constituting the actual appearance order of the second log information. A third log ID 904 indicates the third symbol of each of the plurality of software log information constituting the actual appearance order of the third log information. The fourth log ID 905 indicates the fourth symbol of each of a plurality of pieces of software log information constituting the actual order of appearance of the fourth log information. The serviceman, for example, refers to the log information comparison table 900 to identify the cause of the operation of the image forming apparatus 100 that may be defective.

Next, log acquisition processing executed by the image forming apparatus 100 will be described with reference to FIG. 11 . FIG. 11 is a flow chart showing log acquisition processing executed by the image forming apparatus 100 .

Step S32: The identifying unit 1426 identifies an operation of the image forming apparatus 100 that may have a problem.

Step S34: The operation control unit 1425 controls the operation of the image forming apparatus 100 so as to reproduce the operation specified in step S32. As a result, the image forming apparatus 100 reproduces the action specified in step S32.

Step S36: The third acquisition unit 126 acquires third log information indicating the operation status of the image forming apparatus 100 that reproduced the operation in step S34.

Next, details of log acquisition processing executed by the image forming apparatus 100 will be described with reference to FIGS. 12 to 14. FIG. FIG. 12 is a flowchart showing the first stage of detailed log acquisition processing. FIG. 13 is a flowchart showing the middle stage of detailed log acquisition processing. FIG. 14 is a flowchart showing the latter stage of detailed log acquisition processing.

Step S50: The identifying unit 1426 identifies an operation of the image forming apparatus 100 that may have a problem.

Step S55: The operation control unit 1425 determines whether or not the operation identified by the identification unit 1426 in step S50 is the printing operation. If an affirmative determination (Yes) is made in step S55, that is, if the operation specified by the specifying unit 1426 is the printing operation, the process proceeds to step S60. On the other hand, if a negative determination (No) is made in step S55, that is, if the operation specified by the specifying unit 1426 is not the printing operation, the process proceeds to step S120.

Step S60: The operation control section 1425 controls the operation of the image forming section 110 so that the charging section 110c charges the peripheral surface of the photosensitive drum 110b. As a result, the charging section 110c charges the peripheral surface of the photoreceptor drum 110b.

Step S65: The operation control section 1425 controls the operation of the image forming section 110 so that the exposure section 110a forms an electrostatic latent image on the photosensitive drum 110b. As a result, the exposure unit 110a forms an electrostatic latent image on the photosensitive drum 110b.

Step S70: The operation control section 1425 controls the operation of the image forming section 110 so that the developing section 110d supplies toner to the photosensitive drum 110b. As a result, the developing section 110d supplies toner to the photosensitive drum 110b to develop the electrostatic latent image. Here, in steps S60 to S70, the operation control unit 1425 controls the conveying unit 114 so as not to convey the sheet S from the cassette 112a toward the image forming unit 110. FIG.

Step S75: The third acquisition unit 126 acquires third log information indicating the operational status of the image forming apparatus 100 that reproduced the printing operations in steps S60 to S70.

Step S80: The determination unit 1424 compares the actual appearance order of the second log information with the actual appearance order of the third log information acquired in step S75.

Step S85: The determination unit 1424 determines whether or not the actual appearance order of the second log information is similar to the actual appearance order of the third log information acquired in step S75, based on the comparison result in step S80. judge. If a negative determination (No) is made in step S85, that is, if the actual appearance order of the second log information and the actual appearance order of the third log information do not approximate each other, the process ends (returns). On the other hand, if an affirmative determination (Yes) is made in step S85, that is, if the actual appearance order of the second log information and the actual appearance order of the third log information are close to each other, the process proceeds to step S90.

Step S90: The operation control unit 1425 changes the operating conditions for the printing operation specified by the specifying unit 1426 in step S50.

Step S95: The operation control section 1425 controls the operation of the image forming section 110 so that the charging section 110c charges the peripheral surface of the photosensitive drum 110b based on the operating conditions changed in step S90. As a result, the charging section 110c charges the circumferential surface of the photoreceptor drum 110b based on the operating conditions changed in step S90.

Step S100: The operation control section 1425 controls the operation of the image forming section 110 so that the exposure section 110a forms an electrostatic latent image on the photosensitive drum 110b based on the operating conditions changed in step S90. As a result, the exposure unit 110a forms an electrostatic latent image on the photosensitive drum 110b based on the operating conditions changed in step S90.

Step S105: The operation control section 1425 controls the operation of the image forming section 110 so that the developing section 110d supplies toner to the photosensitive drum 110b based on the operating conditions changed in step S90. As a result, the developing unit 110d develops the electrostatic latent image by supplying toner to the photosensitive drum 110b based on the operating conditions changed in step S90. Here, in steps S95 to S105, the operation control unit 1425 controls the conveying unit 114 so as not to convey the sheet S from the cassette 112a toward the image forming unit 110. FIG.

Step S110: The fourth acquisition unit 128 acquires fourth log information indicating the operation status of the image forming apparatus 100 that reproduced the printing operation by changing the operation conditions in steps S95 to S105.

Step S115: The operation control unit 1425 determines whether the fourth acquisition unit 128 has acquired a predetermined number of patterns of fourth log information. If the determination in step S115 is affirmative (Yes), that is, if the fourth acquisition unit 128 acquires a predetermined number of pieces of fourth log information, the process ends (returns). On the other hand, if a negative determination (No) is made in step S115, that is, if the fourth acquisition unit 128 has not acquired the predetermined number of patterns of fourth log information, the process proceeds to step S90. Then, the operation control unit 1425 changes the operation conditions and causes the image forming apparatus 100 to reproduce the printing operation again.

Step S120: The motion control unit 1425 determines whether or not the motion specified by the specifying unit 1426 in step S50 is a warm-up motion. If an affirmative determination (Yes) is made in step S120, that is, if the motion specified by the specifying unit 1426 is the warm-up motion, the process proceeds to step S125. On the other hand, if a negative determination (No) is made in step S120, that is, if the operation specified by the specifying unit 1426 is not the warm-up operation or the printing operation, the process ends (returns).

Step S125: The operation control section 1425 controls the operation of the image forming apparatus 100 so as to reproduce the warm-up operation. As a result, image forming apparatus 100 reproduces the warm-up operation.

Step S130: The third acquisition unit 126 acquires third log information indicating the operation status of the image forming apparatus 100 that reproduced the warm-up operation in step S125.

Step S135: The determination unit 1424 compares the actual appearance order of the second log information with the actual appearance order of the third log information acquired in step S130.

Step S140: The determination unit 1424 determines whether or not the actual appearance order of the second log information is similar to the actual appearance order of the third log information acquired in step S130, based on the comparison result in step S135. judge. If a negative determination (No) is made in step S140, that is, if the actual appearance order of the second log information and the actual appearance order of the third log information do not approximate each other, the process ends (returns). On the other hand, if an affirmative determination (Yes) is made in step S140, that is, if the actual appearance order of the second log information is close to the actual appearance order of the third log information, the process proceeds to step S145.

Step S145: The operation control unit 1425 changes the operating conditions for the warm-up operation of the image forming apparatus 100 specified by the specifying unit 1426 in step S50.

Step S150: The operation control section 1425 controls the operation of the image forming apparatus 100 to reproduce the warm-up operation based on the operating conditions changed in step S145. As a result, image forming apparatus 100 reproduces the warm-up operation based on the operating conditions changed in step S145.

Step S155: The fourth acquisition unit 128 acquires fourth log information indicating the operation status of the image forming apparatus 100 that reproduced the warm-up operation by changing the operation conditions in step S150.

Step S160: The operation control unit 1425 determines whether the fourth acquisition unit 128 has acquired a predetermined number of patterns of fourth log information. If an affirmative determination (Yes) is made in step S160, that is, if the fourth acquisition unit 128 acquires the predetermined number of patterns of fourth log information, the process ends (returns). On the other hand, if a negative determination (No) is made in step S160, that is, if the fourth acquisition unit 128 has not acquired the predetermined number of patterns of fourth log information, the process proceeds to step S145. Then, the operation control unit 1425 changes the operation conditions and causes the image forming apparatus 100 to reproduce the warm-up operation again.

As described above, according to the image forming apparatus 100 of the present embodiment, it is possible to more comprehensively grasp defects.

The embodiments of the present invention have been described above with reference to the drawings. However, the present invention is not limited to the above-described embodiments, and can be implemented in various aspects without departing from the gist of the present invention. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above embodiments. For example, some components may be omitted from all components shown in the embodiments. Furthermore, components across different embodiments may be combined as appropriate. In order to make the drawings easier to understand, the drawings schematically show each component mainly, and the length, number, spacing, etc. of each component shown are different from the actual ones due to the convenience of drawing. In some cases. Further, the shape, dimensions, etc. of each component shown in the above embodiment are examples and are not particularly limited, and various modifications are possible within a range that does not substantially deviate from the effects of the present invention.

(1) In the above description, the image forming apparatus 100 is electrophotographic, but the present embodiment is not limited to this. The image forming apparatus 100 may be of another type. For example, the image forming apparatus 100 may be of an inkjet type.

(2) As described with reference to FIG. 12, according to the present embodiment, in step S70, the operation control unit 1425 controls the image forming unit 110 so that the developing unit 110d supplies toner to the photosensitive drum 110b. controlled the operation of However, when the image forming unit 110 reproduces the printing operation, the developing unit 110d does not develop the electrostatic latent image formed on the photoreceptor drum 110b as long as the electrostatic latent image is formed on the photoreceptor drum 110b. It's good.

(3) As described with reference to FIG. 2, according to the present embodiment, when the operation specified by the specifying unit 1426 is the printing operation, the operation control unit 1425 directs the image forming unit 110 from the cassette 112a. By controlling the conveying unit 114 so as not to convey the sheet S, the image forming apparatus 100 reproduces the printing operation. However, the operation control unit 1425 may cause the conveying unit 114 to convey the sheet S and cause the image forming apparatus 100 to reproduce the printing operation. In this case, the reproducibility of the printing operation is higher than that of reproducing the printing operation without forming an image on the sheet S. As a result, the fourth log information that makes it easier to identify the cause of the malfunction of the image forming apparatus 100 can be acquired.

The present invention relates to an image forming apparatus and has industrial applicability.

REFERENCE SIGNS LIST 100 image forming apparatus 110 image forming section 110a exposure section 110b photoreceptor drum (image carrier)
110c charging unit 110d developing unit 112a cassette 114 conveying unit 122 first acquiring unit 124 second acquiring unit 126 third acquiring unit 128 fourth acquiring unit 130 communication unit 140 control unit 142 processing unit 1421a first symbolizing unit 1421b second symbol conversion unit 1421c third encoding unit 1422 learning unit 1423 inference unit 1424 determination unit 1425 operation control unit 1426 identification unit 144 storage unit 150 input/output unit 156 audio output unit

Claims (6)

An image forming apparatus,
a first acquisition unit that acquires first log information indicating an operation status of the image forming apparatus;
a first encoding unit that assigns a first symbol uniquely corresponding to the first log information;
a learning unit that inputs the first symbol to a learning model using a neural network and causes the learning model to learn a first probability distribution that indicates the appearance probability of the first log information;
a second acquisition unit that indicates the operation status of the image forming apparatus and acquires second log information having the same format as the first log information;
a second encoding unit that assigns a second symbol uniquely corresponding to the second log information;
inputting the second symbol to the learning model, causing the learning model to output a second probability distribution indicating the appearance probability of the second log information, and based on the second probability distribution, calculating the second log information an inference unit that infers an inference appearance order;
a determination unit that determines whether or not there is a possibility of malfunction of the image forming apparatus based on a comparison between the inferred order of appearance of the second log information and the actual order of appearance of the second log information;
a specifying unit that specifies an operation of the image forming apparatus that may have a problem when the determining unit determines that the image forming apparatus may have a problem;
an operation control unit that controls the operation of the image forming apparatus so as to reproduce the operation specified by the specifying unit;
and a third obtaining unit that obtains third log information indicating the operation status of the image forming apparatus when the image forming apparatus reproduces the operation identified by the identifying unit. an image forming unit that forms an image on a sheet;
a cassette that accommodates the sheet;
a conveying unit that conveys the sheet from the cassette toward the image forming unit;
The image forming unit
an image carrier;
a charging unit that charges the image carrier;
an exposure unit that exposes the charged image carrier to form an electrostatic latent image on the image carrier;
including
When the operation specified by the specifying unit is an operation for the image forming unit to form an image on the sheet, the operation control unit prevents the sheet from being conveyed from the cassette toward the image forming unit. 2. The image forming apparatus according to claim 1, further comprising: controlling said conveying section during operation, and controlling said image forming section such that said exposing section forms said electrostatic latent image on said image carrier. the image forming unit includes a developing unit that supplies a developer to the exposed image carrier to develop the electrostatic latent image;
When the operation specified by the specifying unit is an operation for the image forming unit to form an image on the sheet, the operation control unit causes the developing unit to supply the developer to the image carrier. 3. The image forming apparatus according to claim 2, wherein said image forming unit is controlled by said image forming unit. further comprising a fourth acquisition unit;
The determining unit determines whether or not the actual order of appearance of the second log information and the actual order of appearance of the third log information are similar,
When the determining unit determines that the actual order of appearance of the second log information and the actual order of appearance of the third log information are close to each other, the operation control unit determines the operating condition of the operation specified by the specifying unit. changing the operation of the image forming apparatus to reproduce the operation specified by the specifying unit;
The fourth acquisition unit acquires fourth log information indicating an operation status of the image forming apparatus when reproducing the operation identified by the identification unit by changing the operation condition. Item 4. The image forming apparatus according to any one of Item 3. further comprising a third encoding unit that assigns a third symbol uniquely corresponding to the third log information;
The determination unit compares the actual appearance order of the second symbols of the second log information with the actual appearance order of the third symbols of the third log information, and determines the actual appearance of the second log information. 5. The image forming apparatus according to claim 4, wherein it is determined whether or not the order and the actual appearance order of the third log information are similar. 6. The image forming apparatus according to claim 1, wherein said neural network includes a recursive neural network.

Images