Input device, control method for input device and information processing device

Abstract

To provide an input device, a method for controlling the input device, and an information processing device that accurately detect wrong key pressing.SOLUTION: An input device 10 includes: a keyboard 110 having a sensor 110a that senses the manner in which a key is pressed; a control unit 120 including a correction detection unit 121 that detects a correction operation that is an operation for correcting input contents with the keyboard 110, an operation information acquisition unit 122 that acquires before-correction operation information including the manner in which the key is pressed and information about the key pressed before the detection of the correction operation, and after-correction operation information that is information about the key pressed after the detection of the correction operation, and a model generation unit 123 that generates a learned model 131 for inferring erroneous input by using the before-correction operation information and the after-correction operation information; a storage unit 130 that stores the learned model; and a display unit 140.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an input device, a control method for an input device, and an information processing device. [Background technology]

[0002] In desk work, typing speed affects work efficiency. Typing errors on the keyboard interrupt input and result in lost time for correcting the characters. Patent Document 1 discloses a technology that, when an incorrect string is input, associates the original string with the corrected string and stores them in memory. When the same error is made a predetermined number of times, the incorrect string is automatically corrected to the corrected string. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2003-308316 Summary of the Invention [Problem to be solved by the invention]

[0004] However, if the character string before correction becomes a target for correction after a predetermined number of mistakes, there is a risk that it will be automatically corrected by mistake, even if it is a meaningful character string that the user intentionally entered.

[0005] In one aspect, the present invention aims to provide a technique for detecting keystroke errors with high accuracy. [Means for solving the problem]

[0006] In order to achieve the above object, the present invention employs the following configuration.

[0007] A first aspect of the present disclosure is an input device including: a keyboard having a sensor that detects how keys are pressed; a correction detection unit that detects a correction operation that is an operation for correcting content input via the keyboard; an operation information acquisition unit that acquires pre-correction operation information including information on keys pressed before the correction operation is detected and how the keys are pressed; and post-correction operation information that is information on keys pressed after the correction operation is detected; and a generation unit that generates a trained model that infers erroneous inputs using the pre-correction operation information and the post-correction operation information. The input device can generate a trained model that accurately infers keystroke errors by learning operation information including not only key information but also the physical way the keys are pressed.

[0008] The trained model may be a model that has been trained to output correction candidates when operation information including information on pressed keys and how the keys are pressed is input, if a character string corresponding to the key information is an incorrect input. The input device can determine an incorrect input and obtain correction candidates by inputting the character string entered by the user and the key depression amounts corresponding to each character into the trained model.

[0009] The generation unit may generate the trained model using pre-correction operation information and post-correction operation information by a specific user. The input device can generate the trained model corresponding to the user's unique typing error habits by learning information about the specific user's erroneous inputs.

[0010] The generation unit may generate the trained model using pre-correction operation information and post-correction operation information of a plurality of users. By learning the input information, it is possible to generate a trained model that can handle common mistakes.

[0011] The correction operation may be an operation of pressing the Backspace key or the Delete key. The input device can detect that the user has performed an operation to correct the input content by pressing a specific key.

[0012] The keyboard may include a switch inside the key, and the sensor may measure the amount of depression or pressure of the switch. The input device can detect how the key is pressed by measuring the amount of depression or pressure of the switch, and generate a trained model that infers erroneous inputs by taking into account the physical way the key is pressed.

[0013] Each key on the keyboard may have multiple switches, and the operation information acquisition unit may acquire the key depression position based on the order in which the multiple switches are pressed or the amount of depression of the switches. By providing multiple switches on each key, the keyboard can acquire information about which part of the key the user pressed.

[0014] The input device may include a keyboard having a sensor that detects how the keys are pressed, a determination unit that inputs information about the pressed keys and operation information including how the keys are pressed into a trained model that infers erroneous inputs and determines whether a character string corresponding to the key information is an erroneous input, an acquisition unit that acquires correction candidates for the character string determined to be an erroneous input, and a display unit that presents the correction candidates. The input device can accurately detect erroneous inputs and present appropriate correction candidates to the user.

[0015] The acquisition unit may acquire the correction candidates using a trained model that has been trained to output correction candidates when operation information is input. The input device can acquire appropriate correction candidates by using the trained model that has been trained by machine learning.

[0016] The input device may further include a storage unit that stores pre-correction operation information, including information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information, which is information on keys pressed after the correction operation is detected, in association with each other, and the acquisition unit may acquire, from the storage unit, a character string in the post-correction operation information associated with the pre-correction operation information for the same character string as the character string determined to be an erroneous input. The input device can acquire, as a correction candidate, a character string that has been actually corrected by the user in the past.

[0017] The display unit may present correction candidates together with the character string determined to be an input error. By presenting the correction candidates together with the input content, the user can check the correction candidates and then decide whether or not to correct the input character string.

[0018] The acquisition unit may acquire a plurality of correction candidates, and the display unit may present the plurality of correction candidates in a manner that allows the user to select one. By presenting the plurality of correction candidates, the user can easily correct the input content to the intended one.

[0019] The display unit may delete the character string determined to be an input error and present a correction candidate. By directly correcting the character string determined to be an input error with the correction candidate, the user can reduce the effort required for correction.

[0020] The display unit may highlight the correction in at least one of the character string determined to be an input error and the correction candidate. The highlighting may be performed by changing the character color or adding a highlight. By highlighting the correction, the user can easily find the part where he or she made a typing error. It can be grasped.

[0021] The keyboard may disable user operations for a predetermined period of time after the determination unit determines that the character string is an incorrect input, thereby preventing the user from continuing to input data while the incorrect input exists and reducing the effort required to delete the input data before realizing the error.

[0022] The key information may include at least one of the input character string and key layout. The character string is not limited to Japanese characters, but may also include alphanumeric characters, symbols, etc. By using a trained model that has learned the input character string information and key layout, the input device can accurately detect typing errors made by the user.

[0023] The key pressing method may include information on at least one of the following: the amount of key pressing, the key pressing position, and the time interval between key inputs. By using a trained model that has learned the physical key pressing method, the input device can detect erroneous inputs while taking into account the habits of the physical key pressing method.

[0024] The key information may include information on character strings of words or phrases extracted by syntactic analysis. By using a trained model that has learned meaningful character strings of words or phrases, the input device can detect erroneous input in units of appropriately segmented character strings.

[0025] The input device may further include a timer that measures the time interval between key inputs. By measuring the time interval between key inputs using the timer and generating a trained model that has learned the time interval between key inputs, the input device can accurately detect erroneous inputs in which the input order has been changed.

[0026] A second aspect of the present invention is an information processing device comprising: a keyboard having a sensor that detects how keys are pressed; a correction detection unit that detects a correction operation, which is an operation for correcting content input via the keyboard; an operation information acquisition unit that acquires pre-correction operation information, which includes information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information, which is information on keys pressed after the correction operation is detected; a generation unit that generates a trained model that infers erroneous input using the pre-correction operation information and the post-correction operation information; a determination unit that inputs the operation information, including information on the pressed keys and how the keys are pressed, into the trained model and determines whether a character string corresponding to the key information is an erroneous input; an acquisition unit that acquires correction candidates for the character string determined to be an erroneous input; and a display unit that presents the correction candidates.

[0027] A third aspect of the present invention is a control method for an input device equipped with a keyboard having a sensor that detects how the keys are pressed, the control method for an input device including: a correction detection step of detecting a correction operation, which is an operation for correcting content input via the keyboard; an operation information acquisition step of acquiring pre-correction operation information, which includes information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information, which is information on keys pressed after the correction operation is detected; and a generation step of generating a trained model that infers erroneous inputs, using the pre-correction operation information and the post-correction operation information.

[0028] The method may also be a control method for an input device equipped with a keyboard having a sensor that detects how the keys are pressed, including a determination step of inputting information about the pressed keys and operation information including how the keys are pressed into a trained model that infers erroneous input, and determining whether or not a character string corresponding to the key information is an erroneous input, an acquisition step of acquiring correction candidates for the character string determined to be an erroneous input, and a display step of presenting the correction candidates.

[0029] The present invention also provides a program for implementing the method or a non-uniform program for implementing the program. It can also be regarded as a recording medium that records information over time. The above means and processes can be combined with each other to the greatest extent possible to constitute the present invention. [Effects of the Invention]

[0030] According to the present invention, keystroke errors can be detected with high accuracy. [Brief explanation of the drawings]

[0031] [Figure 1] FIG. 1 is a diagram illustrating an application example of an input device. [Figure 2] FIG. 2 is a block diagram illustrating the configuration of the input device. [Figure 3] FIG. 3 is a diagram illustrating a switch that detects key depression. [Figure 4]FIG. 4 is a flowchart illustrating the learning process. [Figure 5] FIG. 5 is a flowchart illustrating the inference process. [Figure 6] FIG. 6 is a diagram illustrating a plurality of switches that detect key depressions. [Figure 7] FIG. 7 is a diagram for explaining how to obtain the depression amounts of a plurality of switches. [Figure 8] FIG. 8 is a diagram illustrating an erroneous input in which the order of keys to be input is reversed. [Figure 9] FIG. 9 is a block diagram illustrating the configuration of an input device according to the third modification. DETAILED DESCRIPTION OF THE INVENTION

[0032] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to one aspect of the present invention will now be described with reference to the accompanying drawings.

[0033] <Application example> FIG. 1 is a diagram illustrating an example application of an input device. The input device includes a keyboard and the like, and determines whether a key has been pressed incorrectly (incorrect input) based on information about the pressed key and how the key was pressed. The information about the pressed key is, for example, information about the input character string. In the following description, the character string includes numbers, symbols, and the like. The way the key is pressed is, for example, the amount of pressure applied to the key. When inputting characters, if an adjacent key is pressed by mistake, the amount of pressure applied to the erroneously pressed key may be weaker. Therefore, by determining an incorrect input by taking the amount of pressure into consideration, the input device can accurately detect a keystroke.

[0034] First, we will explain how to generate a trained model (dictionary) used to detect input errors. When an input device detects a correction operation such as pressing the Backspace key, it acquires information about the character strings entered before and after the correction, as well as information about the amount of depression of each key.

[0035] Figure 1 shows an example of "KIKAI" being mistakenly typed as "JIKAI." The character string before correction is "JIKAI." If the standard keystroke amount is "2" and the lower keystroke amount is "1," the keystroke amount for each key in the character string "JIKAI" before correction is "12222." The character string after correction is "KIKAI," and the keystroke amount for each key is "22222." Before correction, the "J" was mistakenly pressed for the "K" next to it, and it can be seen that the keystroke amount for the "J" was less than the standard.

[0036] The input device generates a trained model for inferring input errors through machine learning such as deep learning using the pre-correction character string, the depression amounts of each key, and the corrected character string. The trained model may be a model that infers (outputs) whether or not a character string to be judged entered by a user is an input error, or may be a model that infers a correction candidate character string when the character string to be judged is an input error. The input device generates the trained model using the pre-correction character string, the depression amounts of each key, and the corrected character string as training data.

[0037] Next, we will explain how to determine whether an input is incorrect using the generated trained model. The input device inputs information about the input character string and information about the amount of depression of each key (operation information to be determined) into the trained model, and determines whether the input character string is incorrect based on the inference result of the trained model. It is determined whether or not

[0038] For example, if the character string to be judged is "JIKAI" and the amount of depression of each key is "12222," the trained model will infer that the character string to be judged is an input error. On the other hand, if the user intentionally inputs the character string "JIKAI," the amount of depression of each key will be "22222," and the trained model will infer that the input character string is not an input error.

[0039] As described above, the input device can detect keystroke errors with greater accuracy by taking into account the physical key pressing method, such as the key pressing amount. The input device is, for example, a keyboard, and may also be a computer (information processing device) equipped with a display device such as a monitor. The key pressing method is not limited to the key pressing amount, but may also include the key pressing position and the time interval between key inputs. Furthermore, the key pressing amount is not limited to two levels, but may be three or more levels.

[0040] <Embodiment> The device configuration of the input device 10 will be described with reference to Fig. 2. Fig. 2 is a block diagram illustrating the configuration of the input device 10. The input device 10 includes a keyboard 110, a control unit 120, a storage unit 130, and a display unit 140.

[0041] The keyboard 110 is an operating member that accepts operations from a user, and is, for example, a keyboard. Detection of the amount of key depression will now be described with reference to FIG. 3. FIG. 3 is a diagram illustrating a switch 111 that detects key depression. The switch 111 is provided inside the key top and detects key depression. The keyboard 110 includes a sensor 110a that can measure the amount of depression of the switch 111 (the height to which the switch 111 is depressed) or the pressure with which the switch 111 is depressed, as a way of pressing the key. The amount of key depression may be determined based on the pressure with which the switch 111 is depressed.

[0042] The control unit 120 includes a correction detection unit 121 for executing the learning process, an operation information acquisition unit 122, and a model generation unit 123. The control unit 120 also includes an erroneous input determination unit 124 and a candidate acquisition unit 125 for executing the inference process.

[0043] In the learning process, the correction detection unit 121 detects a correction operation, which is an operation for correcting the content input by the user via the keyboard 110. The correction operation is, for example, an operation of pressing a key for correcting the input content, such as the Backspace key or the Delete key. The correction operation may also be an operation of selecting a character string before correction with a pointing device such as a mouse and inputting a character string after correction.

[0044] The operation information acquisition unit 122 acquires pre-correction operation information including information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information which is information on keys pressed after the correction operation is detected.

[0045] The key information is, for example, information on the input character string, and the character string before correction can be, for example, a character string deleted by pressing the Backspace key, etc. In the example of Fig. 1, after "JIKAI" is input, the Backspace key is pressed five times to delete "JIKAI", and input for correction is started, so that the operation information acquisition unit 122 acquires "JIKAI" as the character string before correction.

[0046] In addition, if only the character to be corrected is corrected, the character string before correction can be a character string that includes the character to be manipulated and the character input after it. For example, if "JIKAI" is input, If, after "J" is entered, "J" is deleted and "K" is entered, the operation information acquisition unit 122 can acquire "JIKAI" as the character string before correction, including "IKAI" that was entered after "J".

[0047] For example, if the pre-correction character string is deleted by repeatedly pressing the Backspace key, the corrected character string can be a character string of the same number of characters as the deleted characters from the character string input after the deletion. In the example of Fig. 1, if five characters of "JIKAI" are deleted, the operation information acquisition unit 122 acquires the five characters of "KIKAI" input after the deletion as the corrected character string.

[0048] Also, if "JIKAI" is entered, then the "J" is deleted and a "K" is entered, the operation information acquisition unit 122 may acquire "KIKAI," which is a string of five characters identical to the pre-correction string "JIKAI," as the corrected string.

[0049] The pre-correction character string and the post-correction character string may be information on character strings in units of words or clauses extracted by syntactic analysis of a sentence (natural language) input by the user. Note that the pre-correction character string and the post-correction character string are not limited to Japanese character strings and may include alphanumeric characters, symbols, etc. Even if the incorrectly input character is not at the beginning of a word, the operation information acquisition unit 122 can acquire the pre-correction character string and the post-correction character string in appropriate units by syntactically analyzing the input character string.

[0050] The key pressing method is, for example, the amount of key pressing described in Fig. 1. The amount of key pressing is not limited to two levels and may be three or more levels. The operation information acquisition unit 122 may store the acquired pre-correction operation information and post-correction operation information in the storage unit 130 in association with each other.

[0051] The model generation unit 123 (corresponding to the "generation unit") generates a trained model 131 that infers an erroneous input using the pre-correction operation information and the post-correction operation information. The model generation unit 123 generates the trained model 131 by training, for example, the pre-correction character string "JIKAI" and the push amount "12222" that are the pre-correction operation information, and the corrected character string "KIKAI" that is also the pre-correction operation information.

[0052] Furthermore, the model generation unit 123 may determine whether to learn pre-correction operation information and post-correction operation information as training data based on the degree of coincidence between the pre-correction character string and the post-correction character string. For example, when "BOKUWA" is corrected to "WATASHIWA," it is considered that the input of "BOKUWA" is not an input error and that the user intentionally corrected it to "WATASHIWA." If the degree of coincidence between the pre-correction character string and the post-correction character string is equal to or less than a predetermined threshold, the model generation unit 123 may exclude these character strings from the learning target.

[0053] The degree of match between the pre-correction character string and the corrected character string can be, for example, the ratio of the number of characters in the pre-correction character string that match the corrected character string to the number of characters in the pre-correction character string. Specifically, since "BOKUWA" has six characters and the number of characters that match "WATASHIWA" is two, the degree of match between "BOKUWA" and "WATASHIWA" is 2 / 6 (approximately 33%). Furthermore, since "JIKAI" has five characters and the number of characters that match "KIKAI" is four, the degree of match between "JIKAI" and "KIKAI" is 4 / 5 (80%). If the predetermined threshold is set to 60%, the model generation unit 123 excludes the pre-correction and post-correction character strings "BOKUWA" and "WATASHIWA" from the learning targets.

[0054] In this embodiment, it is assumed that the model generation unit 123 generates the trained model 131 by learning operation information by a specific user who uses the input device 10. When multiple users use the input device 10, the model generation unit 123 may, for example, Alternatively, the user may be identified using account information of the user, and a trained model 131 may be generated for each user.

[0055] In the inference process, the erroneous input determination unit 124 (corresponding to the "determination unit") inputs operation information including information on the pressed key and how the key was pressed into the trained model 131 generated by the model generation unit 123, and determines whether the character string corresponding to the key information is an erroneous input.

[0056] The erroneous input determination unit 124 extracts a character string to be determined as an erroneous input as key information, and acquires the key depression amount corresponding to each character as the key pressing method. The erroneous input determination unit 124 inputs the extracted character string and the acquired key depression amount into the trained model 131, and determines whether the character string to be determined is an erroneous input.

[0057] A method for extracting a character string to be judged as an erroneous input will be described. Each time a character is input, the erroneous input judgment unit 124 acquires a character string including a predetermined number of characters (for example, 10 characters) that were input last, and inputs the last character string of 2, 3, ..., 10 characters as a character string to be judged together with the amount of depression into the trained model 131. The predetermined number is not limited to 10 characters, and may be set to a number of 2 or more characters taking into consideration the load of the erroneous input judgment process.

[0058] The erroneous input determination unit 124 may also analyze the information of the input character string by breaking it down into character strings on a word-by-word or phrase-by-phrase basis, and extract the character string input immediately before as a character string to be determined as an erroneous input. The erroneous input determination unit 124 inputs the extracted character string and the key depression amount corresponding to each character into the trained model 131 as operation information. The erroneous input determination unit 124 determines whether the extracted character string is an erroneous input based on the inference result of the trained model 131.

[0059] The candidate acquisition unit 125 (corresponding to the "acquisition unit") acquires correction candidates for a character string determined to be an incorrect input. The candidate acquisition unit 125 acquires correction candidates, for example, using a trained model 131 that has been trained to output correction candidates when the character string to be judged and the depression amounts of each key (operation information) are input. In the example of FIG. 1, in response to the input character string being "JIKAI" and operation information in which the depression amounts of each key are "12222," the trained model 131 outputs "KIKAI" as a correction candidate.

[0060] The candidate acquisition unit 125 is not limited to acquiring correction candidates using the trained model 131, and may acquire correction candidates from post-correction operation information stored in the storage unit 130 during the learning process. Specifically, the candidate acquisition unit 125 acquires, as correction candidates, from the storage unit 130, character strings of post-correction operation information associated with pre-correction operation information for the same character string as the character string determined to be an erroneous input by the erroneous input determination unit 124.

[0061] The storage unit 130 stores pre-correction operation information, which includes information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information, which includes information on keys pressed after the correction operation is detected and how the keys are pressed, in association with each other. The pre-correction operation information and the post-correction operation information are acquired by the operation information acquisition unit 122. The storage unit 130 also stores the trained model 131 generated by the model generation unit 123.

[0062] The trained model 131 is a trained model for inferring erroneous input. For example, the trained model 131 uses user operation information (key information and key pressing method) as input data to infer whether a character string to be judged corresponding to the key information is an erroneous input. Specifically, the trained model 131 outputs the probability (reliability) that the character string to be judged is an erroneous input. In this case, the erroneous input judgment unit 124 judges that the input character string is an erroneous input when the probability that the character string to be judged is an erroneous input is equal to or greater than a predetermined value (e.g., 80%). It is possible.

[0063] The trained model 131 may also be generated to infer a correction candidate character string when a character string corresponding to key information is an incorrect input. In this case, the trained model 131 uses user operation information as input data and outputs a correction candidate character string and its reliability. The trained model 131 may output multiple correction candidates and their respective reliability.

[0064] For example, if a correction candidate is output and the reliability of the correction candidate is equal to or greater than a predetermined value (e.g., 60%), the incorrect input determination unit 124 can determine that the character string to be determined is an incorrect input. Also, if a correction candidate is not output or the reliability of the correction candidate is less than a predetermined value, the incorrect input determination unit 124 can determine that the character string to be determined is not an incorrect input.

[0065] The display unit 140 is a display or the like, and displays (presents) the correction candidates acquired by the candidate acquisition unit 125. The display unit 140 deletes the character string determined to be an incorrect input, and automatically corrects it to the acquired correction candidate.

[0066] Furthermore, the display unit 140 may present correction candidates together with the character string determined to be an incorrect input. By presenting correction candidates instead of automatically correcting the character string, the user can check the correction candidates and then decide whether or not to correct the input character string.

[0067] When the candidate acquisition unit 125 acquires multiple correction candidates, the display unit 140 may present the acquired multiple correction candidates in a selectable manner for the user. The user can select the correction candidate that is closest to the character string that the user intended to input from the multiple correction candidates.

[0068] The display unit 140 may highlight the corrections in at least one of the original character string and the correction candidates. The display unit 140 can highlight the corrections by, for example, changing the character color or adding a highlight.

[0069] Note that each functional unit in FIG. 2 may or may not be an individual piece of hardware. The functions of two or more functional units may be realized by common hardware. Each of multiple functions of one functional unit may be realized by separate hardware. Two or more functions of one functional unit may be realized by common hardware. Furthermore, each functional unit may or may not be realized by hardware. For example, the device may have a processor and a memory in which a control program is stored. Then, the functions of at least some of the functional units of the device may be realized by the processor reading and executing the control program from the memory.

[0070] (Learning process) The learning process will be described with reference to Fig. 4. The learning process shown in Fig. 4 is started and ended by, for example, an instruction from the user.

[0071] In S101, the keyboard 110 waits for a key input from the user. In S102, the correction detection unit 121 determines whether or not a correction operation has been detected in the key input from the user. A correction operation is an operation of pressing the Backspace key or the like. If a correction operation has been detected, the process proceeds to S103. If a correction operation has not been detected, the process returns to S101.

[0072] In S103, the operation information acquisition unit 122 acquires operation information before and after correction. The operation information before correction is the character string before correction and the amount of pressing of the key corresponding to each character. The character string before correction is, for example, a character string that has been continuously deleted by using the Backspace key or the like, or is a string containing the corrected character and the character that was entered after it.

[0073] The post-correction operation information is a character string after correction. The post-correction character string is, for example, a character string input after a correction operation is detected, and may be a character string having the same number of characters as the character string before correction.

[0074] 1, the operation information acquiring unit 122 acquires the character string "JIKAI" and the depression amount "12222" as pre-correction operation information, and the operation information acquiring unit 122 acquires the character string "KIKAI" as post-correction operation information.

[0075] In S104, the model generation unit 123 generates a trained model 131 that infers erroneous inputs through machine learning using the pre-correction operation information and post-correction operation information acquired in S103. The model generation unit 123 collects information on character strings entered before and after a correction operation and information on the amount of depression of each key while the user is using the keyboard, and can generate a trained model 131 that has learned the tendency of the user to make mistakes.

[0076] The input device 10 may perform the learning process while the input device 10 is set to a learning mode, which is a mode for generating the trained model 131. The learning mode can be switched on and off by the user.

[0077] (inference processing) The inference process will be described with reference to Fig. 5. The inference process shown in Fig. 5 starts and ends in response to an instruction from the user, for example.

[0078] In S201, the keyboard 110 waits for a key input from the user. In S202, the erroneous input determination unit 124 determines whether the character string input from the keyboard 110 is an erroneous input by using the trained model 131 generated in the training process.

[0079] The erroneous input determination unit 124 acquires the input character string to be determined as key information and acquires the key depression amount as the key depression method. The erroneous input determination unit 124 inputs the acquired character string and the key depression amount of each character into the trained model 131, and can determine whether the input character string is an erroneous input. If the character string to be determined is an erroneous input, the process proceeds to S203. If the character string to be determined is not an erroneous input, the process returns to S201.

[0080] If it is determined in S202 that the character string to be determined is an incorrect input, the keyboard 110 may disable input from the user for a predetermined time. By disabling input from the keyboard 110, the user can avoid continuing input while there is an incorrect input, and can reduce the effort required to delete content that was input before realizing the error. The predetermined time is, for example, the time from when the input is determined to be an incorrect input to when the user realizes the error and stops key input.

[0081] In S203, the candidate acquisition unit 125 acquires correction candidates based on the character string determined to be an incorrect input in S202. The candidate acquisition unit 125 may acquire correction candidates by inputting the character string determined to be an incorrect input into the trained model 131, or may acquire, from the storage unit 130, a character string of post-correction operation information corresponding to the character string determined to be an incorrect input, as a correction candidate.

[0082] In S204, the display unit 140 presents the correction candidates acquired in S203 on the display. When the display unit 140 automatically converts the incorrectly input character string into the correction candidates, the user can reduce the effort required for correction. When the display unit 140 presents correction candidates for the incorrectly input character string, the user can confirm whether or not the correction is the intended correction before making the correction. can be converted into candidates.

[0083] The input device 10 may execute the inference process while set to an inference mode, which is a mode for detecting and correcting an erroneous input. The inference mode can be switched on and off by the user.

[0084] According to the above embodiment, the input device 10 can accurately detect keystroke errors by using the trained model 131 that has learned the user's tendency to make keystroke errors through machine learning. Furthermore, the input device 10 can improve typing speed by converting a character string determined to be an incorrect input into a correction candidate or presenting the correction candidate.

[0085] <Variation 1> In the above embodiment, the input device 10 learns the tendency of typing errors of a specific user to generate the trained model 131. In Modification 1, the input device 10 learns the tendency of typing errors of a plurality of users to generate the trained model 131. That is, the input device 10 generates the trained model by using pre-correction operation information and post-correction operation information input by a plurality of users.

[0086] The input device 10 can generate a trained model 131 that corresponds to common mistakes by learning information on various user input errors without distinguishing between users. A common mistake is, for example, when trying to press "S" but pressing the adjacent "A" to the left with a smaller pressing force of "1" than normal. Even before the user has learned their own mistakes, the user can efficiently correct keystroke errors by using the trained model 131 that corresponds to common mistakes.

[0087] <Variation 2> In the above embodiment, the keyboard 110 has one switch 111 for detecting key depression inside the key top of each key, as shown in Fig. 3. In the second modification, each key has multiple switches.

[0088] Fig. 6 is a diagram illustrating a plurality of switches that detect key presses. In the example of Fig. 6, a plurality of switches 111a to 111d that detect key presses are provided inside the key top. By providing a plurality of switches 111a to 111d for each key, keyboard 110 can obtain information on the key press position, i.e., which part of the key surface is pressed.

[0089] In the second modification, it is sufficient for the switches 111a to 111d to be able to detect whether they have been pressed, and the input device 10, for example, detects which of the multiple switches reacted first. The input device 10 can acquire the position of the switch that reacted first among the switches 111a to 111d as the key depression position. The model generation unit 123 generates a trained model 131 that has learned the key depression position as a key depression method, and the input device 10 can determine an erroneous input by taking the key depression position into consideration.

[0090] The input device 10 may acquire the order in which multiple switches provided on each key are pressed as the way of pressing the key. The model generation unit 123 generates a trained model 131 that has learned the order in which multiple switches are pressed as the way of pressing the key, and the input device 10 can determine an erroneous input by taking into account the order in which the multiple switches are pressed.

[0091] The input device 10 may also acquire the depression amounts of each of a plurality of switches as a key depression. The model generation unit 123 generates a trained model 131 that has learned the depression amounts of the plurality of switches as a key depression, and the input device 10 can determine an erroneous input by taking into account the depression amounts of the plurality of switches.

[0092] Acquisition of the depression amounts of multiple switches will be described with reference to Fig. 7. Fig. 7 shows an example in which "KIKAI" has been mistyped as "JIKAI." The character string before correction is "JIKAI." If the standard depression amount is "2," a depression amount less than the standard is "1," and no depression is "0," then the depression amounts of "J" for switches 111a to 111d are "1," "1," "0," and "0," respectively.

[0093] 7 shows that switch 111b located on the side of the "K" key and switch 111a located on the side of the "I" key, which is input after "K," have been pressed, although less than the standard amount. On the other hand, switches 111c and 111d, which are located away from the "K" and "I" keys, have not detected any pressing.

[0094] The model generation unit 123 generates a trained model 131 that has learned the depression amounts of multiple switches as key pressing methods, and the input device 10 is able to determine erroneous input by taking into account the depression amounts of multiple switches.

[0095] Using the generated trained model 131, the erroneous input determination unit 124 can determine that "JIKAI" is an erroneous input of "KIKAI" when the depression amounts for the switches 111a to 111d are "1", "1", "0", and "0". On the other hand, if the user intentionally inputs the character string "JIKAI", it is expected that the depression amounts for the switches 111a to 111d will be "2", "2", "2", and "2". In this case, the erroneous input determination unit 124 can determine that the input of "JIKAI" (with depression amounts of "2", "2", "2", "2") is not an erroneous input.

[0096] In the second modification, the operation information acquisition unit 122 acquires, as a key pressing method, the key pressing position, the pressing order of the multiple switches, the pressing amounts of the multiple switches, etc. The model generation unit 123 generates a trained model 131 that has been trained as a key pressing method using at least one of the pieces of information acquired by the operation information acquisition unit 122. By taking into account the key pressing position, the pressing order of the multiple switches, the pressing amounts of the multiple switches, etc., the input device 10 can accurately detect keystroke errors.

[0097] Furthermore, the number and arrangement of the multiple switches are not limited to the example in Figure 7. As many switches as can be arranged inside one key may be arranged along each side as in Figure 7, or may be arranged at each corner or in the center.

[0098] <Variation 3> In the above embodiment, the input device 10 detects the amount of key depression as the way a key is pressed. In Modification 2, the input device 10 detects the key depression position, etc. as the way a key is pressed. In Modification 3, the input device 10 further measures the time interval between key inputs and uses this to determine whether an input is incorrect.

[0099] Fig. 8 is a diagram illustrating an erroneous input in which the order of keys entered is reversed. The time interval between key inputs with the right hand and left hand becomes short, and the user may accidentally input the key that was intended to be pressed next with the left hand before the key that is to be pressed with the right hand. In the example of Fig. 8, the user may mistakenly input "A" → "K" in the order that they intended to input "K" → "A."

[0100] If the user mistakenly inputs "A" → "K" in that order, it is expected that the time interval between the keystrokes of "A" and "K" will be approximately simultaneous. In this way, if the time interval between the keystrokes differs from (for example, is shorter than) the user's average time interval between keystrokes, it is possible that the user made a typing error. On the other hand, if the time interval between the keystrokes is approximately the same as the user's average time interval between keystrokes, it is considered that the user intentionally input "A" → "K" in that order.

[0101] Fig. 9 is a block diagram illustrating the configuration of an input device according to Modification 3. The input device 10 includes a timer 150 that measures the time interval between key inputs. Since the configuration other than the timer 150 is the same as the configuration shown in Fig. 2, only the differences from the above embodiment and each modification will be described.

[0102] The operation information acquisition unit 122 acquires the time intervals between key inputs measured by the timer 150 as the way the keys are pressed. The model generation unit 123 generates a trained model 131 that has learned the time intervals between key inputs, and the input device 10 is able to determine an erroneous input by taking the time intervals between key inputs into consideration. Therefore, the input device 10 can accurately detect an erroneous input even when the input order of keys input by the left and right hands is reversed.

[0103] <Other> The above-described embodiment and each modification merely exemplify the configuration of the present invention. The present invention is not limited to the specific embodiments described above, and various modifications are possible within the scope of the technical concept.

[0104] In the above embodiment, the learning process and the inference process are described as being executed separately, but they may be executed in parallel. By continuously executing the learning process, the input device 10 can generate a trained model 131 that can determine erroneous input with higher accuracy.

[0105] Furthermore, in the above embodiment, an example has been described in which the input device 10 generates the trained model 131. However, the trained model 131 may be generated by an external device capable of communicating with the input device 10. In this case, the external device acquires the pre-correction operation information and post-correction operation information acquired by the operation information acquisition unit 122 online from the input device 10. The external device can generate the trained model 131 in the same way as the model generation unit 123, using the pre-correction operation information and post-correction operation information acquired from the input device 10.

[0106] The key information may also include information about key layout, such as whether or not the key for the character before correction (including keys other than characters) and the key for the character after correction are adjacent to each other.

[0107] The operation information acquisition unit 122 acquires key layout information as key information. The model generation unit 123 generates a trained model 131 that has learned the key layout information, and the input device 10 is able to determine erroneous inputs by taking the key layout information into consideration. Because keystroke errors tend to occur between adjacent keys, by learning the key layout information, the input device 10 can accurately detect erroneous inputs.

[0108] <Appendix 1> a keyboard (110) equipped with a sensor (110a) for detecting how the keys are pressed; a correction detection unit (121) that detects a correction operation that is an operation for correcting the input content via the keyboard (110); Pre-correction operation information including information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information which is information on the keys pressed after the correction operation is detected. an operation information acquisition unit (122) that acquires information; a generation unit (123) that generates a trained model (131) that infers an erroneous input using the pre-correction operation information and the post-correction operation information; An input device (10) comprising:

[0109] <Appendix 2> a keyboard (110) equipped with a sensor (110a) for detecting how the keys are pressed; a determination unit (124) that inputs information about a pressed key and operation information including how the key is pressed into a trained model (131) that infers an erroneous input, and determines whether a character string corresponding to the key information is an erroneous input; an acquisition unit (125) that acquires correction candidates for the character string determined to be an incorrect input; a display unit (140) that presents the correction candidates; An input device (10) comprising:

[0110] <Appendix 3> a keyboard (110) equipped with a sensor (110a) for detecting how the keys are pressed; a correction detection unit (121) that detects a correction operation that is an operation for correcting the input content via the keyboard (110); an operation information acquisition unit (122) that acquires pre-correction operation information including information on keys pressed before detecting the correction operation and how the keys are pressed, and post-correction operation information that is information on the keys pressed after detecting the correction operation; a generation unit (123) that generates a trained model (131) that infers an erroneous input using the pre-correction operation information and the post-correction operation information; a determination unit (124) that inputs information about a pressed key and operation information including how the key is pressed into the trained model (131) and determines whether a character string corresponding to the key information is an incorrect input; an acquisition unit (125) that acquires correction candidates for the character string determined to be an incorrect input; a display unit (140) that presents the correction candidates; An information processing device (10) comprising:

[0111] <Appendix 4> A method for controlling an input device (10) having a keyboard (110) equipped with a sensor (110a) that detects how a key is pressed, comprising: a correction detection step (S102) for detecting a correction operation for correcting the input content via the keyboard (110); an operation information acquisition step (S103) of acquiring pre-correction operation information including information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information which is information on the keys pressed after the correction operation is detected; A generation step (S104) of generating a trained model (131) for inferring an erroneous input using the pre-correction operation information and the post-correction operation information; A method for controlling an input device (10) comprising:

[0112] <Appendix 5> A method for controlling an input device (10) having a keyboard (110) equipped with a sensor (110a) that detects how a key is pressed, comprising: A determination step (S202) of inputting information about a pressed key and operation information including how the key was pressed into a trained model (131) that infers an erroneous input, and determining whether or not a character string corresponding to the key information is an erroneous input; an acquisition step (S203) of acquiring correction candidates for the character string determined to be an incorrect input; a display step (S204) of presenting the correction candidates; A method for controlling an input device (10) comprising: [Explanation of symbols]

[0113] 10: input device, 110: keyboard, 110a: sensor, 111: switch, 111a: switch, 111b: switch, 111c: switch, 111d: switch, 120: control unit, 121: correction detection unit, 122: operation information acquisition unit, 123: model generation unit, 124: erroneous input determination unit, 125: candidate acquisition unit, 130: storage unit, 131: model, 140: display unit, 150: timer

Claims

1. a keyboard having a sensor that detects how the keys are pressed; a correction detection unit that detects a correction operation that is an operation for correcting the input content via the keyboard; an operation information acquisition unit that acquires pre-correction operation information including information on keys pressed before detecting the correction operation and how the keys are pressed, and post-correction operation information that is information on the keys pressed after detecting the correction operation; a generation unit that generates a trained model for inferring an erroneous input using the pre-correction operation information and the post-correction operation information; Equipped with The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. Input device.

2. The trained model is a model that has been trained to infer correction candidates when operation information including information on pressed keys and how the keys are pressed is input and a character string corresponding to the key information is an incorrect input. The input device according to claim 1 .

3. The generation unit generates the trained model using the pre-correction operation information and the post-correction operation information by a specific user.

3. The input device according to claim 1 or 2.

4. The generation unit generates the trained model using the pre-correction operation information and the post-correction operation information of a plurality of users. The input device according to claim 1 .

5. The correction operation is an operation of pressing a Backspace key or a Delete key. The input device according to claim 1 .

6. each key of the keyboard comprises a plurality of the switches; the operation information acquisition unit acquires key depression positions based on the depression order of the plurality of switches or the depression amounts of the switches; The input device according to claim 1 .

7. a keyboard having a sensor that detects how the keys are pressed; a determination unit that inputs information about a pressed key and operation information including how the key is pressed into a trained model that infers an input error, and determines whether a character string corresponding to the key information is an input error; and an acquisition unit that acquires correction candidates for the character string determined to be an input error; a display unit that presents the correction candidates; Equipped with The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. Input device.

8. the acquisition unit acquires the correction candidates using the trained model that has been trained to output the correction candidates when the operation information is input. The input device according to claim 7.

9. a storage unit that stores pre-correction operation information, which includes information on the keys pressed before the correction operation is detected and how the keys are pressed, in association with post-correction operation information, which is information on the keys pressed after the correction operation is detected; the acquiring unit acquires, from the storage unit, as the correction candidate, a character string in the post-correction operation information associated with the pre-correction operation information for the same character string as the character string determined to be an incorrect input. The input device according to claim 7.

10. the display unit presents the correction candidate together with the character string determined to be an input error.

10. An input device according to any one of claims 7 to 9.

11. the acquisition unit acquires a plurality of the correction candidates, the display unit presents a plurality of the correction candidates in a manner that allows a user to select one of them; 11. The input device according to claim 7.

12. the display unit deletes the character string determined to be an incorrect input and presents the correction candidate.

10. An input device according to any one of claims 7 to 9.

13. the display unit highlights a correction in at least one of the character string determined to be an input error and the correction candidate.

13. An input device according to any one of claims 7 to 12.

14. the keyboard disables operations from the user for a predetermined time after the determination unit determines that the character string is an incorrect input.

14. An input device according to any one of claims 7 to 13.

15. The key information includes at least one of an input character string and a key layout.

15. An input device according to any one of claims 1 to 14.

16. The key pressing method includes at least one of information on the key pressing amount, the key pressing position, and the time interval of the key input.

16. An input device according to any one of claims 1 to 15.

17. The key information includes information on character strings of words or phrases extracted by syntactic analysis.

17. An input device according to any one of claims 1 to 16.

18. Further provided with a timer for measuring the time interval between key inputs; 18. An input device according to any one of claims 1 to 17.

19. a keyboard having a sensor that detects how the keys are pressed; a correction detection unit that detects a correction operation that is an operation for correcting the input content via the keyboard; an operation information acquisition unit that acquires pre-correction operation information including information on keys pressed before detecting the correction operation and how the keys are pressed, and post-correction operation information that is information on the keys pressed after detecting the correction operation; a generation unit that generates a trained model for inferring an erroneous input using the pre-correction operation information and the post-correction operation information; a determination unit that inputs information about a pressed key and operation information including a manner in which the key is pressed into the trained model and determines whether a character string corresponding to the information about the key is an incorrect input; an acquisition unit that acquires correction candidates for the character string determined to be an input error; a display unit that presents the correction candidates; Equipped with The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. Information processing device.

20. A method for controlling an input device having a keyboard with a sensor that detects how a key is pressed, comprising: a correction detection step of detecting a correction operation for correcting the input content via the keyboard; an operation information acquisition step of acquiring pre-correction operation information including information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information which is information on the keys pressed after the correction operation is detected; a generation step of generating a trained model for inferring an erroneous input using the pre-correction operation information and the post-correction operation information; Including, The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. A method for controlling an input device.

21. A method for controlling an input device having a keyboard with a sensor that detects how a key is pressed, comprising: a determination step of inputting information about the pressed keys and operation information including how the keys were pressed into a trained model that infers erroneous inputs, and determining whether or not a character string corresponding to the key information is an erroneous input; an acquisition step of acquiring a correction candidate for the character string determined to be an incorrect input; a display step of presenting the correction candidates; Including, The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. A method for controlling an input device.

22. A control program for an input device having a keyboard with a sensor that detects how the keys are pressed, a correction detection step of detecting a correction operation for correcting the input content via the keyboard; an operation information acquisition step of acquiring pre-correction operation information including information on keys pressed before the correction operation is detected and how the keys are pressed, and post-correction operation information which is information on the keys pressed after the correction operation is detected; a generation step of generating a trained model for inferring an erroneous input using the pre-correction operation information and the post-correction operation information; Including, The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. Input device control program.

23. A control program for an input device having a keyboard with a sensor that detects how the keys are pressed, a determination step of inputting information about the pressed keys and operation information including how the keys were pressed into a trained model that infers erroneous inputs, and determining whether or not a character string corresponding to the key information is an erroneous input; an acquisition step of acquiring a correction candidate for the character string determined to be an incorrect input; a display step of presenting the correction candidates; Including, The keyboard has switches inside the keys, The sensor measures the depression amount or depression pressure of the switch. Input device control program.

Images