Relationship determination device
The relationship determination device simplifies the determination of text relationships by using a classification model to output probability information, addressing the complexity of existing methods and enhancing efficiency and accuracy.
Patent Information
- Application Number
- JP2023079180
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-12
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
Smart Images

Figure 2024163496000001_ABST
Abstract
Description
[Technical field]
[0001] One aspect of the present disclosure relates to a relationship determination device that determines a relationship between texts each composed of one or more words. [Background technology]
[0002] The following Patent Document 1 discloses a discussion situation understanding support device that performs morphological analysis on messages posted via email and BBS functions on the Web, and performs pattern matching between the analysis results and patterns for determining actions that are composed of predetermined action types (questions, confirmations, requests, doubts, answers, agreements, etc.), matching patterns (patterns expressed using parts of speech), and matching examples, to identify the type of action for each message and determine the relationships between messages. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2009-26137 A Summary of the Invention [Problem to be solved by the invention]
[0004] In the above-mentioned discussion situation grasping support device, in order to find the relationship between messages, complex processing is required, such as performing morphological analysis and then performing pattern matching with a complexly constructed action determination pattern. Therefore, it is desired to more easily determine the relationship between texts. [Means for solving the problem]
[0005] A relationship determination device according to one aspect of the present disclosure includes a determination unit that obtains first probability information, which is probability information output by inputting a first text that is a first text, and second probability information, which is probability information output by inputting a second text that is a second text, for a classification model that outputs probability information regarding the probability that, when a text consisting of one or more words is input, the text will be classified into each of a plurality of predetermined labels, and determines the relationship between the first text and the second text based on the obtained first probability information and second probability information.
[0006] In this aspect, the relationship between the first text and the second text is determined based on first probability information output by inputting the first text to the classification model and second probability information output by inputting the second text to the classification model, which makes it easier to determine the relationship between the texts. Effect of the Invention
[0007] According to one aspect of the present disclosure, relationships between texts can be determined more easily. [Brief description of the drawings]
[0008] [Figure 1] FIG. 2 is a diagram illustrating an example of a functional configuration of a relationship determination device according to an embodiment. [Diagram 2] 5 is a flowchart illustrating an example of a determination process executed by a relationship determination device according to the embodiment. [Diagram 3] FIG. 13 is a diagram showing example English composition questions. [Figure 4] FIG. 4 is a diagram showing example answers by examinees to the example questions shown in FIG. 3. [Diagram 5] FIG. 1 illustrates an example of problem formulation. [Figure 6] FIG. 13 is a diagram illustrating an example of learning a classification model. [Figure 7] 1 is a diagram showing an example of determination by a relationship determination device according to an embodiment; [Figure 8]10 is a flowchart illustrating another example of the determination process executed by the relationship determination device according to the embodiment. [Figure 9] FIG. 2 is a diagram illustrating an example of a hardware configuration of a computer used in the relationship determination device according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0009] Hereinafter, the embodiments of the present disclosure will be described in detail with reference to the drawings. In the description of the drawings, the same elements are given the same reference numerals, and duplicated descriptions will be omitted. In addition, the embodiments of the present disclosure in the following description are specific examples of the present invention, and the present invention is not limited to these embodiments unless otherwise specified to limit the present invention.
[0010] Fig. 1 is a diagram showing an example of the functional configuration of a relationship determination device 1 according to an embodiment. The relationship determination device 1 is a computer device that determines the relationship between texts each consisting of one or more words. As shown in Fig. 1, the relationship determination device 1 includes a storage unit 10, an acquisition unit 11, a learning unit 12, an input unit 13 (input unit), and a determination unit 14 (determination unit).
[0011] Each functional block of the relationship determination device 1 is assumed to function within the relationship determination device 1, but is not limited to this. For example, some of the functional blocks of the relationship determination device 1 may function in a computer device different from the relationship determination device 1 and connected to the relationship determination device 1 through a network, while appropriately transmitting and receiving information to and from the relationship determination device 1. Also, some of the functional blocks of the relationship determination device 1 may be omitted, multiple functional blocks may be integrated into one functional block, or one functional block may be decomposed into multiple functional blocks.
[0012] Hereinafter, each function of the relationship determination device 1 shown in FIG. 1 will be described.
[0013] The storage unit 10 stores any information used in calculations and the like in the relationship determination device 1, and the results of the calculations in the relationship determination device 1. The information stored by the storage unit 10 may be referred to by each function of the relationship determination device 1 as appropriate.
[0014] The storage unit 10 may store, for example, a text, a first text, a second text, a first text group, a second text group, a pair, an extracted pair, a (trained) classification model, a plurality of predetermined labels, probability information, first probability information, or second probability information, as described below.
[0015] The acquisition unit 11 acquires any information used for calculations and the like in the relationship determination device 1. The acquisition unit 11 may acquire information from another device via a network using a communication device 1004 described below, or may acquire information input by a user of the relationship determination device 1 using an input device 1005. The acquisition unit 11 may cause the storage unit 10 to store the acquired information, or may output the acquired information to the learning unit 12, the input unit 13, and the determination unit 14.
[0016] The learning unit 12 learns the classification model, and the learned classification model is stored by the storage unit 10. When a text consisting of one or more words is input, the classification model outputs probability information regarding the probability that the text is classified into each of a plurality of predetermined labels. The classification model performs so-called multi-label classification.
[0017] Multi-label classification is a classification problem in which multiple labels are assigned (given) to a single object (text). Generally, in multi-label classification, first, the object to be labeled is input into a model, and a probability value (for example, a real number between "0" and "1") for each label is calculated. Next, for each label probability value calculated by the model, if the value exceeds a predetermined threshold, the label is assigned (the flag for the label is set to "1" (true)), and if the value is below the threshold, the label is not assigned (the flag for the label is set to "0" (false)). This binary classification is then performed for all labels.
[0018] The text may be, for example, a single word, two or more words, a phrase, a clause, a phrase, a sentence, a paragraph, or a paragraph.
[0019] A label is one or more words. For example, a label may be a single word, two or more words, a phrase, a clause, or a phrase.
[0020] The plurality of predetermined labels may be composed of, for example, four labels, "Women", "Children", "Pressure", and "Lifestyle". The plurality of predetermined labels may be composed of, for example, four labels, "Society", "Politics", "Economy", and "History". The plurality of predetermined labels is not limited to four labels, and may be composed of any number of labels equal to or greater than two.
[0021] Specific examples of probability information include the probability that the label “society” is assigned “0.9” (90%), the probability that the label “politics” is assigned “0.2” (20%), the probability that the label “economy” is assigned “0.1” (10%), and the probability that the label “history” is assigned “0.1” (10%).
[0022] The learning of the classification model by the learning unit 12 is performed by existing techniques such as machine learning, similar to model learning in multi-label classification.
[0023] The input unit 13 inputs any text to the classification model stored by the storage unit 10. When the input unit 13 inputs text, the classification model outputs probability information regarding the probability that the text is classified into each of a plurality of predetermined labels. The probability information output by the classification model is used by the determination unit 14 described below.
[0024] The input unit 13 may input a first text, which is a first text, to the classification model. The input unit 13 may input a second text, which is a second text, to the classification model. In the embodiment, the first text and the second text are assumed to be different, but may be the same.
[0025] The first text may be a word or phrase. The first text may be a word or phrase specified in advance (by a human). The second text may be a phrase or sentence. The second text may be a phrase or sentence created by a human.
[0026] Each of the plurality of predetermined labels (which are classified into the classification model) may be different from at least one of the first text or the second text (which is input to the classification model). More specifically, each of the plurality of predetermined labels may be different from the first text, each of the plurality of predetermined labels may be different from the second text, or each of the plurality of predetermined labels may be different from the first text and / or the second text.
[0027] Any of a plurality of predetermined labels (as classification destinations of the classification model) may be the same as at least one of the first text or the second text (as input to the classification model). More specifically, any of a plurality of predetermined labels may be the same as the first text, any of a plurality of predetermined labels may be the same as the second text, or any of a plurality of predetermined labels may be the same as the first text and / or the second text.
[0028] The input unit 13 may extract one or more pairs each consisting of a text included in a first text group consisting of a plurality of texts and a text included in a second text group consisting of a plurality of texts, and for each extracted pair, the text from the first text group may be input as a first text and the text from the second text group may be input as a second text to the classification model in sequence.
[0029] The input unit 13 may extract one or more pairs each consisting of a text included in a first text group consisting of a plurality of texts, each being a word or a phrase, and a text included in a second text group consisting of a plurality of texts, each being a phrase or a sentence, and for each of the extracted pairs, the text from the first text group may be input as a first text and the text from the second text group as a second text to the classification model in sequence.
[0030] When the first text and the second text of each extraction pair are sequentially input by the input unit 13, the classification model sequentially outputs, for each extraction pair, probability information regarding the first text of the extraction pair and probability information regarding the second text of the extraction pair. The probability information of each extraction pair sequentially output by the classification model is used by the determination unit 14 described below.
[0031] The determination unit 14 acquires first probability information, which is probability information output by inputting a first text, which is a first text, and second probability information, which is probability information output by inputting a second text, which is a second text, into the classification model, and determines the relationship between the first text and the second text based on the acquired first probability information and second probability information.
[0032] The relationship may be, for example, the degree of similarity between the first text and the second text, whether the first text and the second text are the same, whether the first text and the second text are approximately the same, whether the first text and the second text are semantically close, whether the content of the first text contains the content of the second text, whether the content of the second text contains the content of the first text, whether the content of the first text is in line with the content of the second text, or whether the content of the second text is in line with the content of the first text.
[0033] The judgment unit 14 may store the judgment result in the storage unit 10, or may output (transmit) it to another device via a network using the communication device 1004 described below, or may output (display) it to a user of the relationship judgment device 1 using the output device 1006.
[0034] The determination unit 14 may determine the relationship based on the similarity between the first probability information and the second probability information. The determination unit 14 may determine the relationship based on the similarity between the probability of being classified into each of a plurality of predetermined labels indicated by the first probability information and the probability of being classified into each of a plurality of predetermined labels indicated by the second probability information. The similarity is determined using an existing technique such as cosine similarity.
[0035] The determination unit 14 may determine whether or not the content of the first text includes the content of the second text. The determination unit 14 may determine whether or not the content of the second text includes the content of the first text.
[0036] The determination unit 14 may determine the relationship for each extracted pair (based on the probability information for each extracted pair output sequentially by the classification model) and determine the relationship between the first text group and the second text group based on the determination result for each extracted pair.
[0037] The judgment unit 14 may judge, for each extraction pair, whether or not the content of the second text of the extraction pair includes the content of the first text of the extraction pair (based on the probability information of each extraction pair output sequentially by the classification model), and may judge, based on the judgment result of each extraction pair, whether or not the content of the second text group includes the content of the first text group.
[0038] Next, an example of processing executed by the relationship determination device 1 will be described with reference to Fig. 2. Fig. 2 is a flowchart showing an example of the determination processing executed by the relationship determination device 1.
[0039] First, the learning unit 12 learns a classification model and causes the storage unit 10 to store the learned classification model (step S1). Next, the input unit 13 inputs a first text to the classification model stored in S1 (step S2). Next, the determination unit 14 acquires first probability information output by the classification model based on S2 (step S3). Next, the input unit 13 inputs a second text to the classification model stored in S1 (step S4). Next, the determination unit 14 acquires second probability information output by the classification model based on S4 (step S5). Next, the determination unit 14 determines the relationship between the first text and the second text based on the first probability information acquired in S3 and the second probability information acquired in S5 (step S6).
[0040] The order of S2 to S5 may be changed as long as the conditions that S3 is performed after S1 and before S6, and S2 is performed after S5, and S4 is performed after S5 are satisfied. For example, the order may be S1, S2, S4, S3, S5, and S6, or S1, S4, S5, S2, S3, and S6. Also, S1 may be omitted, and an already trained classification model may be used for S2 and S4.
[0041] Hereinafter, a specific embodiment will be used to describe the details of each process of the relationship determination device 1. As a specific embodiment, an example will be described in which the relationship determination device 1 is used to score English sentences written by examinees (people, users) in response to English composition questions in an English proficiency test.
[0042] FIG. 3 is a diagram showing an example of an English composition problem. The example problem shown in FIG. 3 includes a TOPIC and POINTS. Candidates are required to write an essay (in English) about the given TOPIC. In addition, in the essay, they are required to use two of the POINTS to support their answer. In addition, the essay is required to be composed of an introduction, a main body, and a conclusion, and is recommended to be 120 to 150 words long. In the example problem shown in FIG. 3, candidates are required to write an essay about the TOPIC "Is marriage less important today than in the past?" using two of the four POINTS, "Women," "Children," "Pressure," and "Life style." FIG. 4 is a diagram showing an example of a response by a test-taker to the example problem shown in FIG. 3. That is, the example response shown in FIG. 4 is an essay written by a test-taker.
[0043] As described above, in the example question shown in FIG. 3, it is necessary to select a keyword (POINTS) from POINTS and answer the question. The answer (essay) only needs to contain the keyword, and does not need to contain the word or phrase (of POINTS) itself. POINTS is not necessarily a word. In this embodiment, the relationship determination device 1 aims to automatically determine whether or not a keyword is included in an essay written by a test-taker.
[0044] When the problem in this embodiment is formulated, it can be formulated as a multi-label classification problem in which an answer document is divided into multiple units and then multiple labels are assigned. Keywords are used as labels, and a classification model is trained using training data. FIG. 5 is a diagram showing an example of problem formulation. The formulation example shown in FIG. 5 shows a scene in which the classification model is trained so that when "For example, women who are married and have children tend to be unable to maintain their careers.", which is a unit (phrase or sentence) extracted from the essay shown in FIG. 4, is input to the classification model, the corresponding label ("Women" and "Children") is output (multiple labels can also be output).
[0045] However, in the above formulation, the number of labels must be determined in advance before training the classification model. In addition, it is necessary to label not only superficial character strings but also to understand the context in which they are used. Other issues include the lack of training data for the classification model, and the fact that the labels (POINTS) differ depending on the question (not limited to the four, "Women," "Children," "Pressure," and "Life style"). Specifically, (1) there is no training data in the essay domain, (2) the POINTS differ depending on the question, so the number of labels would be enormous if they were to be applied to all questions, and (3) adding a new question would require retraining the classification model.
[0046] In the scoring using the relationship determination device 1, in the learning of the classification model (by the learning unit 12), the classifier is trained with a general multi-label classification data set. In the learning, the POINTS and the labels do not have to match. FIG. 6 is a diagram showing an example of learning of the classification model. In the learning example shown in FIG. 6, learning is performed with a data set consisting of the text "The ex-president announced his candidacy almost four months ago.", which is a general multi-label classification data set, and the correct answer data "label: society, flag: 0", "label: politics, flag: 1", "label: economy, flag: 1" and "label: history, flag: 0". Here, the labels "society", "politics", "economy" and "history" (multiple predetermined labels) are labels that are expected to be used in essays to some extent, which are prepared in advance and fixed. In this way, the multiple predetermined labels may be labels unrelated to POINTS (labels outside the data, labels outside the target data). The correct answer data may be created in advance by a person.
[0047] In scoring using the relationship determination device 1, the determination unit 14 determines (infers) the relationship using the classification model trained as described above. Fig. 7 is a diagram showing an example of determination by the relationship determination device 1. In the determination example shown in Fig. 7, first, the determination unit 14 acquires the first probability information "Label: Society, Probability: 0.9", "Label: Politics, Probability: 0.2", "Label: Economy, Probability: 0.1", and "Label: History, Probability: 0.1" that is output when the input unit 13 inputs "Women", one of the POINTS shown in Fig. 3, as the first text, into the classification model trained as described above. Next, the input unit 13 inputs, as the second text, "For example, women who are married and have children tend to be unable to maintain their careers.", which is a part (sentence) of the essay shown in FIG. 4, and the determination unit 14 acquires the second probability information "label: society, probability: 0.8", "label: politics, probability: 0.1", "label: economy, probability: 0", and "label: history, probability: 0.2". Next, the determination unit 14 compares the similarity between the distribution of the acquired first probability information and the distribution of the second probability information to determine the relationship between the first text and the second text. For example, if the distributions are similar (the similarity meets a predetermined criterion), the determination unit 14 determines that the first text and the second text are semantically close. Also, for example, if the distributions are not similar (the similarity does not meet a predetermined criterion), the determination unit 14 determines that the first text and the second text are not semantically close (distant).
[0048] Next, an example of processing executed by the relationship determination device 1 in this embodiment will be described with reference to Fig. 8. Fig. 8 is a flowchart showing another example of the determination processing executed by the relationship determination device 1.
[0049] First, the learning unit 12 learns using data with general labels to create a classification model, which is then stored by the storage unit 10 (step S10). Next, the examinee writes an essay, which is acquired by the acquisition unit 11 and stored by the storage unit 10 (step S11). Next, the input unit 13 divides the essay stored in S11 into phrases or sentences (step S12). Next, the input unit 13 extracts one or more pairs each consisting of one of the POINTS (prestored by the storage unit 10 or acquired by the acquisition unit 11) and one of the divided essays (in S12) (step S13).
[0050] The following steps S14 to S18 are repeatedly performed for each extracted pair, which is a pair extracted in S13. First, the input unit 13 inputs the POINTS of the extracted pair to the classification model stored in S10 (step S14). Next, the judgment unit 14 acquires the first probability information output from the classification model based on S14 (step S15). Next, the input unit 13 inputs the divided essay of the extracted pair to the classification model stored in S10 (step S16). Next, the judgment unit 14 acquires the second probability information output from the classification model based on S16 (step S17). Next, the judgment unit 14 judges whether or not the content of the POINTS of the extracted pair is included in the content of the divided essay of the extracted pair based on the first probability information acquired in S15 and the second probability information acquired in S17 (step S18).
[0051] After steps S14 to S18 have been performed for each extracted pair, the judgment unit 14 judges whether or not the content of the essay (answered by the examinee) includes the content of POINTS (of the English composition question) based on the judgment results for each extracted pair (step S19).
[0052] S10 may be omitted, and an already trained classification model may be used in S14 and S16. The order of S14 to S17 may be changed as long as the following conditions are satisfied: S15 is performed after S13 and before S18, and S14 is performed after S15, and S17 is performed after S16. After S19, the judgment unit 14 may output the judgment result in S19. The pairs extracted in S13 may be all pairs (pairs of all combinations of POINTS and divided essays), or may be some of the pairs. The repeated process of S14 to S18 may be stopped and proceed to S19 if a predetermined condition is satisfied midway, for example, if it is determined that the contents of two POINTS (required in an English composition question) are included in the contents of the essay (since it is pointless to confirm that the contents of three or more POINTS are included).
[0053] As a result of the above, the relationship determination device 1 can automatically determine whether or not keywords or POINTS (first text or first text group) are included in an essay (second text or second text group) written by an examinee.
[0054] Next, the effects of the relationship determination device 1 according to the embodiment will be described.
[0055] According to the relationship determination device 1, a classification model outputs probability information regarding the probability that a text consisting of one or more words is classified into each of a plurality of predetermined labels when the text is input, and a determination unit 14 is provided that acquires first probability information, which is probability information output by inputting a first text that is a first text, and second probability information, which is probability information output by inputting a second text that is a second text, for the classification model, and determines the relationship between the first text and the second text based on the acquired first probability information and second probability information. With this configuration, the relationship between the first text and the second text is determined based on the first probability information output by inputting the first text to the classification model and the second probability information output by inputting the second text to the classification model. In other words, the relationship between the texts can be determined more easily.
[0056] Furthermore, in the relationship determination device 1, each of the multiple predetermined labels may be different from at least one of the first text or the second text. With this configuration, it is not necessary to match the multiple predetermined labels that are classification destinations by the classification model with the first text or the second text. As a result, for example, a classification model in which multiple predetermined labels consisting of general labels are classification destinations can be used. In other words, the relationship determination device 1 can determine the relationship using a general classification model that is not specialized for the first text or the second text. This makes it easier to determine the relationship between the texts.
[0057] Furthermore, in the relationship determination device 1, any one of the multiple predetermined labels may be the same as at least one of the first text and the second text. With this configuration, a more accurate relationship can be determined.
[0058] Furthermore, in the relationship determination device 1, the determination unit 14 may determine the relationship based on the similarity between the first probability information and the second probability information. With this configuration, it is possible to more accurately determine the relationship based on the similarity between the first probability information and the second probability information.
[0059] Furthermore, in the relationship determination device 1, the determination unit 14 may determine the relationship based on the similarity between the probability of being classified into each of the multiple predetermined labels indicated by the first probability information and the probability of being classified into each of the multiple predetermined labels indicated by the second probability information. With this configuration, it is possible to more accurately determine the relationship based on the similarity between the probability of being classified into each of the multiple predetermined labels indicated by the first probability information and the probability of being classified into each of the multiple predetermined labels indicated by the second probability information.
[0060] Furthermore, in the relationship determination device 1, the determination unit 14 may determine whether or not the content of the first text is included in the content of the second text. With this configuration, it is possible to determine a more specific relationship, that is, whether or not the content of the first text is included in the content of the second text.
[0061] Furthermore, in the relationship determination device 1, the first text may be a word or phrase, and the second text may be a phrase or sentence, and the determination unit 14 may determine whether or not the content of the second text includes the content of the first text. With this configuration, it is possible to determine a more specific relationship, that is, whether or not the content of the first text, which is a word or phrase, is included in the content of the second text, which is a phrase or sentence.
[0062] Furthermore, in the relationship determination device 1, the first text may be a pre-specified word or phrase, and the second text may be a phrase or sentence created by a person, and the determination unit 14 may determine whether or not the content of the second text includes the content of the first text. With this configuration, it is possible to determine a more specific relationship, that is, whether or not the content of the first text, which is a pre-specified word or phrase, is included in the content of the second text, which is a phrase or sentence created by a person.
[0063] The relationship determination device 1 may further include an input unit 13 that extracts one or more pairs each composed of a text included in a first text group composed of a plurality of texts and a text included in a second text group composed of a plurality of texts, and sequentially inputs the text from the first text group as the first text and the text from the second text group as the second text to the classification model for each extracted pair, and the determination unit 14 determines the relationship for each extracted pair and determines the relationship between the first text group and the second text group based on the determination result for each extracted pair. With this configuration, it is possible to determine the relationship between the first text group composed of a plurality of texts and the second text group composed of a plurality of texts.
[0064] The relationship determination device 1 may further include an input unit 13 that extracts one or more pairs each composed of a text included in a first text group composed of a plurality of texts, each of which is a word or phrase, and a text included in a second text group composed of a plurality of texts, each of which is a phrase or sentence, and sequentially inputs the text from the first text group as the first text and the text from the second text group as the second text to the classification model for each extracted pair, and the determination unit 14 may determine whether or not the content of the second text of each extracted pair includes the content of the first text of the extracted pair, and determine whether or not the content of the first text group is included in the content of the second text group based on the determination result of each extracted pair. With this configuration, it is possible to determine whether or not the content of the first text group composed of a plurality of texts, each of which is a phrase or sentence, includes the content of the second text group composed of a plurality of texts, each of which is a word or phrase.
[0065] The relationship determination device 1 executes a method for extracting keywords contained in a document using labels outside the data. The relationship determination device 1 can determine whether an essay is written semantically from the perspective of POINTS. The relationship determination device 1 can automatically determine whether the essay is graded correctly in accordance with the answer specifications (using two of the POINTS).
[0066] The relationship determination device 1 of the present disclosure may have the following configuration.
[0067] [1] A relationship determination device comprising: a determination unit that obtains first probability information, which is the probability information output by inputting a first text, which is a first said text, and second probability information, which is the probability information output by inputting a second text, which is a second said text, for a classification model that outputs probability information regarding the probability that, when a text consisting of one or more words is input, the text will be classified into each of a plurality of predetermined labels, and that determines a relationship between the first text and the second text based on the obtained first probability information and second probability information.
[0068] [2] Each of the plurality of predetermined labels is different from at least one of the first text or the second text; The relationship determination device according to [1].
[0069] [3] Any one of the plurality of predetermined labels is the same as at least one of the first text or the second text. The relationship determination device according to [1].
[0070] [4] the determination unit determines the relationship based on a similarity between the first probability information and the second probability information. The relationship determination device according to any one of [1] to [3].
[0071] [5] the determination unit determines the relationship based on a similarity between a probability of being classified into each of the plurality of predetermined labels indicated by the first probability information and a probability of being classified into each of the plurality of predetermined labels indicated by the second probability information. The relationship determination device according to any one of [1] to [4].
[0072] [6] The determination unit determines whether or not the content of the second text includes the content of the first text. The relationship determination device according to any one of [1] to [5].
[0073] [7] the first text is a word or a phrase; the second text being a phrase or a sentence; The determination unit determines whether or not the content of the second text includes the content of the first text. The relationship determination device according to any one of [1] to [6].
[0074] [8] the first text being a pre-specified word or phrase; the second text being a human-created phrase or sentence; The determination unit determines whether or not the content of the second text includes the content of the first text. The relationship determination device according to any one of [1] to [7].
[0075] [9] an input unit that extracts one or more pairs each composed of one of the texts included in a first text group composed of a plurality of the texts and one of the texts included in a second text group composed of the plurality of the texts, and sequentially inputs the text from the first text group as the first text and the text from the second text group as the second text to the classification model for each of the extracted pairs; the determination unit determines the relationship for each of the extracted pairs, and determines the relationship between the first text group and the second text group based on a determination result for each of the extracted pairs. The relationship determination device according to any one of [1] to [8].
[0076]
[10] an input unit that extracts one or more pairs each composed of one of the texts included in a first text group composed of a plurality of the texts, the texts being words or phrases, and one of the texts included in a second text group composed of a plurality of the texts, the texts being phrases or sentences, and sequentially inputs the text from the first text group as the first text and the text from the second text group as the second text to the classification model for each of the extracted pairs, the determination unit determines, for each of the extraction pairs, whether or not the content of the second text of the extraction pair includes the content of the first text of the extraction pair, and determines, based on a result of the determination for each of the extraction pairs, whether or not the content of the second text group includes the content of the first text group. The relationship determination device according to any one of [1] to [9].
[0077] The block diagrams used to explain the above embodiments show functional blocks. These functional blocks (components) are realized by any combination of at least one of hardware and software. The method of realizing each functional block is not particularly limited. That is, each functional block may be realized by using one device that is physically or logically combined, or may be realized by using two or more devices that are physically or logically separated and directly or indirectly connected (for example, by wire, wirelessly, etc.). The functional blocks may be realized by combining the one device or the multiple devices with software.
[0078] Functions include, but are not limited to, judgment, determination, judgment, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, consideration, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocation, mapping, assignment, etc. For example, a functional block (component) that performs the function of transmission is called a transmitting unit or a transmitter. In either case, as described above, there is no particular limitation on the method of realization.
[0079] For example, the relationship determination device 1 according to an embodiment of the present disclosure may function as a computer that performs processing of the relationship determination method of the present disclosure. Fig. 9 is a diagram showing an example of a hardware configuration of the relationship determination device 1 according to an embodiment of the present disclosure. The relationship determination device 1 described above may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, and the like.
[0080] In the following description, the term "apparatus" may be replaced with a circuit, a device, a unit, etc. The hardware configuration of the relationship determination apparatus 1 may be configured to include one or more of the apparatuses shown in the drawings, or may be configured to exclude some of the apparatuses.
[0081] Each function in the relationship determination device 1 is realized by loading a specific software (program) onto hardware such as a processor 1001 and a memory 1002, causing the processor 1001 to perform calculations, control communication by a communication device 1004, and control at least one of the reading and writing of data in the memory 1002 and the storage 1003.
[0082] The processor 1001 controls the entire computer by running an operating system, for example. The processor 1001 may be configured with a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, a register, etc. For example, the above-mentioned acquisition unit 11, learning unit 12, input unit 13, and determination unit 14 may be realized by the processor 1001.
[0083] Moreover, the processor 1001 reads out a program (program code), a software module, data, etc. from at least one of the storage 1003 and the communication device 1004 into the memory 1002, and executes various processes according to the read out programs. As the program, a program that causes a computer to execute at least a part of the operations described in the above-mentioned embodiment is used. For example, the acquisition unit 11, the learning unit 12, the input unit 13, and the determination unit 14 may be realized by a control program stored in the memory 1002 and operated in the processor 1001, and other functional blocks may be realized in a similar manner. Although the above-mentioned various processes have been described as being executed by one processor 1001, they may be executed by two or more processors 1001 simultaneously or sequentially. The processor 1001 may be implemented by one or more chips. The program may be transmitted from a network via a telecommunication line.
[0084] The memory 1002 is a computer-readable recording medium, and may be configured by at least one of, for example, a Read Only Memory (ROM), an Erasable Programmable ROM (EPROM), an Electrically Erasable Programmable ROM (EEPROM), a Random Access Memory (RAM), etc. The memory 1002 may be called a register, a cache, a main memory (primary storage device), etc. The memory 1002 can store a program (program code), software modules, etc. that are executable to implement a wireless communication method according to an embodiment of the present disclosure.
[0085] Storage 1003 is a computer-readable recording medium, and may be composed of at least one of, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy (registered trademark) disk, a magnetic strip, and the like. Storage 1003 may be referred to as an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, a server, or other suitable medium including at least one of memory 1002 and storage 1003.
[0086] The communication device 1004 is hardware (transmission / reception device) for performing communication between computers via at least one of a wired network and a wireless network, and is also called, for example, a network device, a network controller, a network card, a communication module, etc. The communication device 1004 may be configured to include a high-frequency switch, a duplexer, a filter, a frequency synthesizer, etc., in order to realize at least one of Frequency Division Duplex (FDD) and Time Division Duplex (TDD). For example, the above-mentioned acquisition unit 11, learning unit 12, input unit 13, and determination unit 14, etc. may be realized by the communication device 1004.
[0087] The input device 1005 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that accepts input from the outside. The output device 1006 is an output device (e.g., a display, a speaker, an LED lamp, etc.) that performs output to the outside. The input device 1005 and the output device 1006 may be integrated into one device (e.g., a touch panel).
[0088] In addition, each device such as the processor 1001 and the memory 1002 is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or may be configured using different buses between each device.
[0089] Furthermore, the relationship determination device 1 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA), and some or all of the functional blocks may be realized by the hardware. For example, the processor 1001 may be implemented using at least one of these pieces of hardware.
[0090] Notification of information is not limited to the aspects / embodiments described in this disclosure and may be performed using other methods.
[0091] Each aspect / embodiment described in the present disclosure may be applied to at least one of systems using LTE (Long Term Evolution), LTE-A (LTE-Advanced), SUPER 3G, IMT-Advanced, 4G (4th generation mobile communication system), 5G (5th generation mobile communication system), FRA (Future Radio Access), NR (new Radio), W-CDMA (registered trademark), GSM (registered trademark), CDMA2000, UMB (Ultra Mobile Broadband), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, UWB (Ultra-WideBand), Bluetooth (registered trademark), and other suitable systems, and next-generation systems extended based on these. In addition, a combination of multiple systems (e.g., a combination of at least one of LTE and LTE-A and 5G, etc.) may be applied.
[0092] The order of the steps, sequences, flow charts, etc. of each aspect / embodiment described in this disclosure may be changed unless inconsistent. For example, the methods described in this disclosure present elements of various steps using an example order, and are not limited to the particular order presented.
[0093] The input and output information may be stored in a specific location (e.g., memory) or may be managed using a management table. The input and output information may be overwritten, updated, or added to. The output information may be deleted. The input information may be transmitted to another device.
[0094] The determination may be based on a value represented by a single bit (0 or 1), a Boolean value (true or false), or a numerical comparison (e.g., comparison with a predetermined value).
[0095] Each aspect / embodiment described in this disclosure may be used alone, in combination, or switched according to execution. In addition, notification of specific information (e.g., notification that "X is the case") is not limited to being done explicitly, but may be done implicitly (e.g., not notifying the specific information).
[0096] Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described in the present disclosure. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is intended to be illustrative and does not have any limiting meaning on the present disclosure.
[0097] Software shall be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.
[0098] Additionally, software, instructions, information, etc. may be transmitted or received over a transmission medium. For example, if the software is transmitted from a website, server, or other remote source using wired and / or wireless technologies (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL)), and / or wireless technologies (such as infrared, microwave, etc.), then these wired and / or wireless technologies are included within the definition of transmission media.
[0099] The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.
[0100] In addition, terms explained in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings.
[0101] As used in this disclosure, the terms "system" and "network" are used interchangeably.
[0102] Furthermore, the information, parameters, etc. described in this disclosure may be expressed using absolute values, may be expressed using relative values from a predetermined value, or may be expressed using other corresponding information.
[0103] The names used for the above parameters are not limiting in any way, and furthermore, the formulas etc. using these parameters may differ from those explicitly disclosed in this disclosure.
[0104] As used in this disclosure, the terms "determining" and "determining" may encompass a wide variety of actions. "Determining" and "determining" may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, search, inquiry (e.g., searching in a table, database, or other data structure), ascertaining, and the like. "Determining" and "determining" may also include receiving (e.g., receiving information), transmitting (e.g., sending information), input, output, accessing (e.g., accessing data in a memory), and the like. In addition, "judgment" and "decision" can include considering resolving, selecting, choosing, establishing, comparing, etc., to be a "judgment" or "decision." In other words, "judgment" and "decision" can include considering some action to be a "judgment" or "decision." In addition, "judgment" can be interpreted as "assuming," "expecting," "considering," etc.
[0105] The terms "connected," "coupled," or any variation thereof, refer to any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are "connected" or "coupled" to each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, "connected" may be read as "access." As used in this disclosure, two elements may be considered to be "connected" or "coupled" to each other using at least one of one or more wires, cables, and printed electrical connections, as well as electromagnetic energy having wavelengths in the radio frequency range, microwave range, and optical (both visible and invisible) range, as some non-limiting and non-exhaustive examples.
[0106] As used in this disclosure, the phrase "based on" does not mean "based only on," unless expressly stated otherwise. In other words, the phrase "based on" means both "based only on" and "based at least on."
[0107] Any reference to an element using a designation such as "first," "second," etc., used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient method of distinguishing between two or more elements. Thus, a reference to a first and a second element does not imply that only two elements may be employed or that the first element must precede the second element in some way.
[0108] The "means" in the configuration of each of the above devices may be replaced with "part," "circuit," "device," etc.
[0109] When the terms "include," "including," and variations thereof are used in this disclosure, these terms are intended to be inclusive, similar to the term "comprising." Further, when used in this disclosure, the term "or" is not intended to be an exclusive or.
[0110] In this disclosure, where articles have been added due to translation, such as, for example, a, an, and the in English, the disclosure may include that the nouns following these articles are in the plural form.
[0111] In the present disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "coupled" may also be interpreted in the same way as "different." [Explanation of symbols]
[0112] 1...relationship determination device, 10...storage unit, 11...acquisition unit, 12...learning unit, 13...input unit, 14...determination unit, 1001...processor, 1002...memory, 1003...storage, 1004...communication device, 1005...input device, 1006...output device, 1007...bus.
Claims
1. A relationship determination device comprising: a classification model that outputs probability information regarding the probability that, when text consisting of one or more words is input, the text will be classified into each of a plurality of predetermined labels; a determination unit that obtains first probability information that is output by inputting a first text, which is a first said text, and second probability information that is output by inputting a second text, which is a second said text, and determines the relationship between the first text and the second text based on the obtained first probability information and second probability information.
2. Each of the plurality of predetermined labels is different from at least one of the first text or the second text; The relationship determination device according to claim 1 .
3. Any one of the plurality of predetermined labels is the same as at least one of the first text or the second text. The relationship determination device according to claim 1 .
4. the determination unit determines the relationship based on a similarity between the first probability information and the second probability information. The relationship determination device according to claim 1 .
5. the determination unit determines the relationship based on a similarity between a probability of being classified into each of the plurality of predetermined labels indicated by the first probability information and a probability of being classified into each of the plurality of predetermined labels indicated by the second probability information. The relationship determination device according to claim 1 .
6. The determination unit determines whether or not the content of the second text includes the content of the first text. The relationship determination device according to claim 1 .
7. the first text is a word or a phrase; the second text being a phrase or a sentence; The determination unit determines whether or not the content of the second text includes the content of the first text. The relationship determination device according to claim 1 .
8. the first text being a pre-specified word or phrase; the second text being a human-created phrase or sentence; The determination unit determines whether or not the content of the second text includes the content of the first text. The relationship determination device according to claim 1 .
9. an input unit that extracts one or more pairs each composed of one of the texts included in a first text group composed of a plurality of the texts and one of the texts included in a second text group composed of the plurality of the texts, and sequentially inputs the text from the first text group as the first text and the text from the second text group as the second text to the classification model for each of the extracted pairs; the determination unit determines the relationship for each of the extracted pairs, and determines the relationship between the first text group and the second text group based on a determination result for each of the extracted pairs. The relationship determination device according to claim 1 .
10. an input unit that extracts one or more pairs each composed of one of the texts included in a first text group composed of a plurality of the texts, the texts being words or phrases, and one of the texts included in a second text group composed of a plurality of the texts, the texts being phrases or sentences, and sequentially inputs the text from the first text group as the first text and the text from the second text group as the second text to the classification model for each of the extracted pairs, the determination unit determines, for each of the extraction pairs, whether or not the content of the second text of the extraction pair includes the content of the first text of the extraction pair, and determines, based on a result of the determination for each of the extraction pairs, whether or not the content of the second text group includes the content of the first text group. The relationship determination device according to claim 1 .
Citation Information
Patent Citations
Apparatus and method for supporting obtaining of discussion state
JP2009026137A
