Information processing systems, information processing programs, and storage media

Abstract

This invention provides an information processing system, program, and storage medium capable of obtaining appropriate output when processing information content using artificial intelligence. [Solution] The information processing system 100 is a system for processing information content C, and when information content C is input, it includes an input receiving unit 11 that receives the input of information content C, an information partitioning unit 13 that partitions the information content C so that each partition contains a feature and generates multiple information partitions A1 to A5 containing the features, an information extraction unit 14 that, when extracting information from each of the information partitions A1 to A5, extracts information from the information partitions A1 to A5 so that the total amount of information extracted is less than a reference value Bst and generates multiple extracted information a1 to a5 corresponding to the multiple information partitions A1 to A5, and an information processing instruction unit 15 that executes an instruction to process the information while the multiple extracted information a1 to a5 are combined so that they become one information content.

Description

Technical Field

[0001] The present invention relates to an information processing system for processing information included in information content having a plurality of features, an information processing program for processing the information, and a storage medium storing the program.

Background Art

[0002] Conventionally, as an example of information processing, a technique for summarizing text information is known. In many cases, a summary text is used in order to efficiently understand the content of a large amount of text information. For such generation of a summary text, automatic language processing is suitable. For example, Patent Document 1 discloses a technique for generating a summary text using artificial intelligence.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

[0020] etc. of the specification)

Summary of the Invention

[0004] According to the technique of Patent Document 1 described above, automatic generation of a summary text becomes possible. However, when the amount of information in the text serving as the basis for the summary text is extremely large, it may be difficult to appropriately generate the summary text. This is considered to be caused by the fact that the summarization range is partially biased in the whole text having a large amount of information.

[0005] That is, since the range to be summarized is extremely large, an event may occur in which only a partial and biased part is summarized and the remaining part is not summarized. When the features of the text are expressed in the remaining part, the features are not reflected in the summary text. Therefore, in processing information content using artificial intelligence, information processing that can obtain an appropriate output is desired.

[0006] In view of the above, the object of the present invention is to provide an information processing system, a program to be executed by a computer for information processing, and a storage medium storing the program, which are capable of obtaining appropriate output when processing information content using artificial intelligence. [Means for solving the problem]

[0007] The technical means of the present invention for solving this technical problem is characterized by the following: The information processing system of the present invention is an information processing system for processing information contained in information content that is recognizable by sight and / or hearing and has a plurality of features. The information processing system of the present invention comprises: an input receiving unit that receives the input of information content when the information content is input; an information partitioning unit that partitions the received information content so that each of the features is included, thereby generating a plurality of information partitions containing the features; an information extraction unit that, when extracting information from each of the generated plurality of information partitions, extracts information from the information partitions so that the total amount of information to be extracted is less than a reference value, thereby generating a plurality of extracted pieces of information corresponding to the plurality of information partitions; and an information processing instruction unit that executes an instruction to process the information while the plurality of extracted pieces of information generated are combined so that they become a single information content.

[0008] In the information processing system of the present invention, the information partitioning unit generates at least a first information partition and a second information partition, each having a different amount of information, and the information extraction unit extracts a portion of the information from the first information partition and the second information partition that has a larger amount of information to generate the extracted information, and extracts all of the information from the partition that has a smaller amount of information to generate the extracted information.

[0009] In the information processing system of the present invention, the information partitioning unit generates at least a first information partition and a second information partition, each having a different amount of information, and the information extraction unit generates extracted information by extracting a portion of the information from both the first information partition and the second information partition.

[0010] The information processing system of the present invention further comprises an information acquisition unit that acquires the information content to be input from a database in which a plurality of the information contents are stored.

[0011] In the information processing system of the present invention, the information content includes at least document information, the information partitioning unit generates a plurality of information partitions in the document information, the information extraction unit generates a plurality of extracted information in the document information, and the information processing instruction unit instructs the processing of document summarization as the information processing.

[0012] The information processing system of the present invention further comprises an output display unit that displays and outputs a summary text generated based on the instructions for processing the document summary.

[0013] In the information processing system of the present invention, the information content further includes drawing information, and the output display unit displays and outputs the drawing information in addition to the summary text.

[0014] The present invention provides an information processing program for processing information contained in information content that is recognizable by sight and / or hearing and has multiple features. The present invention provides an information processing program for causing a computer to execute the following when the information content is input: an input receiving means for receiving the input of the information content; an information partitioning means for partitioning the received information content so that each of the features is included, thereby generating a plurality of information partitions containing the features; an information extraction means for extracting information from the plurality of information partitions such that the total amount of information extracted is less than a reference value, thereby generating a plurality of extracted pieces of information corresponding to the plurality of information partitions; and an information processing instruction means for executing an instruction to process the plurality of extracted pieces of information, which have been combined so that they form a single information content. Information processing program.

[0015] The storage medium of the present invention stores the above-mentioned information processing program. [Effects of the Invention]

[0016] According to the present invention, appropriate output can be obtained when processing information content using artificial intelligence. [Brief explanation of the drawing]

[0017] [Figure 1] This is a schematic diagram of an information processing system according to an embodiment of the present invention. [Figure 2] Figure 1 is a diagram illustrating an example of information processing performed by a computer. [Figure 3] Figure 1 is a diagram illustrating an example of information processing performed by a computer. [Figure 4] Figure 1 is a diagram illustrating an example of information processing performed by a computer. [Figure 5]It is a diagram showing a state in which an abstract and drawings are output and displayed on the computer shown in FIG. 1. [Figure 6] It is a flowchart showing a process when text summarization processing is executed as an example of information processing by an information processing system according to an embodiment of the present invention.

Mode for Carrying Out the Invention

[0018] Hereinafter, embodiments of the present invention will be described based on the drawings.

[0019] As shown in FIG. 1, an information processing system 100 according to an embodiment of the present invention is a system for processing information included in information content C. The information processing system 100 includes a computer 10, a server 20, and a database 30. The computer 10, the server 20, and the database 30 can communicate information via an information communication network N.

[0020] The information content C is recognizable visually and / or auditorily and has a plurality of features. The information content C may include, for example, text information, drawing information, audio information, video information, etc. The information content C may be, for example, a patent document (patent gazette, published patent gazette, etc.), an academic document (paper, conference proceedings, etc.), a law-related document (judgment example, law text, etc.), and is not limited thereto. In the present embodiment, it is assumed that the information content C is a patent document including text information and drawing information.

[0021] The computer 10 includes an input reception unit 11, an information acquisition unit 12, an information partitioning unit 13, an information extraction unit 14, an information processing instruction unit 15, and an output display unit 16. The server 20 includes an information processing unit 21. The input reception unit 11, the information acquisition unit 12, the information partitioning unit 13, the information extraction unit 14, the information processing instruction unit 15, the output display unit 16, and the information processing unit 21 are each composed of an electric / electronic circuit, a program stored in a CPU, etc. The computer 10, the server 20, and the server 30 have a communication device (not shown) and can communicate with each other.

[0022] When the information content C is input, the input reception unit 11 receives the input of the information content C. The information content C may be received by the computer 10 via the information communication network N as in this embodiment and recorded in, for example, the memory of the computer 10. Also, the data format at the time of receiving the input may be any format and may be appropriately converted according to the mode of information processing. More specifically, for example, when the text information included in the information content C is image data or PDF data, the input of the data may be received and converted into character data by OCR. Also, when the input reception unit 11 receives PDF data, for example, it may execute extraction of text (character) data from the PDF data. At this time, if it is determined that the extraction of the text data is successful, the extracted text data is used. On the other hand, if it is determined that the extraction of the text data is not successful, the PDF data is used as image data. In this case, the information processing unit 21 in the server 20 described later may perform OCR processing on the image data. Further, when the PDF is partitioned for each page and only the pages corresponding to an information amount smaller than the reference value Bst described later are extracted, only the image data of the extracted pages may be processed by the information processing unit 21 in the server 20.

[0023] The information acquisition unit 12 acquires information content C for input to the computer 10 from the database 30, which stores multiple information content C. That is, information content C for input to the computer 10 is extracted from the database 30 and becomes the subject of information processing. In the database 30, for example, patent documents (i.e., information content C) are stored linked to numbers (publication number, patent number, etc.), dates (filing date, publication date, registration date, etc.), classifications (IPC, FI, etc.), etc. Such patent documents can be extracted from the database 30 by searching the above-mentioned management items or keywords. The extraction and acquisition of information content C may be performed, for example, by accessing the database 30 on the computer 10 and having the above-mentioned management items, keywords, etc. entered by the user.

[0024] As shown in Figures 1 and 2, the information partitioning unit 13 partitions the information content C received by the computer 10 so that each partition contains a feature, thereby generating multiple information partitions containing features. As an example, let's assume that the information content C is a patent publication. The information content C as a patent publication includes text information C1 and drawing information C2. Generally, the text information C1 of a patent publication has headings such as [Claims], [Background Art], [Problems to be Solved by the Invention], [Means for Solving the Problems], and [Modes for Carrying Out the Invention]. Each piece of text information C1 has different features from each other so that the description corresponding to each heading can be disclosed.

[0025] Based on this, in this embodiment, the information partitioning unit 13 may partition the text information C1 corresponding to each "heading". More specifically, the text information C1 of [Claims], [Background Art], [Problems to be Solved by the Invention], [Means for Solving the Problems], and [Embodiments for Carrying Out the Invention] is partitioned, and information partitions A1, A2, A3, A4, and A5 are generated. These information partitions A1, A2, A3, A4, and A5 will include the characteristics of each "heading". Assume that the information partitions A1, A2, A3, A4, and A5 have corresponding information amounts B1, B2, B3, B4, and B5 respectively. Here, assume that the magnitude relationship of the information amounts is B3 < B2 < B1 = B4 < B5. The information amount increases as the number of characters in the text information C1 increases. Thus, the information partitioning unit 13 may generate information partitions having different information amounts respectively.

[0026] When the information extraction unit 14 extracts information from a plurality of information partitions generated by the information partitioning unit 13, it extracts information from the information partitions so that the total amount of the extracted information is smaller than a reference value, and generates a plurality of pieces of extracted information corresponding to the plurality of information partitions. For example, when the total amount of the information amounts B1 to B5 is larger than the reference value Bst of the information amount, the information extraction unit 14 may extract information so that the total amount of the information extracted from the information partitions A1 to A5 is smaller than the reference value Bst. Various modes of information extraction in the information extraction unit 14 can be adopted. As an example, the following three patterns will be described.

[0027] Referring to FIG. 2, the first pattern of information extraction in the information extraction unit 14 will be described. In the first pattern, the information extraction unit 14 extracts a part of the information from the information partition having a larger information amount and generates the extracted information, and extracts all of the information from the information partition having a smaller information amount and generates the extracted information among the information partitions A1 to A5 having different information amounts respectively. Thereby, the total amount of the information extracted from the information partitions A1 to A5 becomes smaller than the reference value Bst. Here, the reference value Bst of the information amount may be, for example, the upper limit of the information amount that can be appropriately processed by the information processing unit 21 of the server 20.

[0028] More specifically, for example, the text information C1 of the information content C is divided into five parts, and information compartments A1 to A5 are generated. Using the information amount obtained by dividing the reference value Bst of the information amount into five equal parts as a threshold value, the threshold value and the information amounts B1 to B5 may be compared respectively. Among the information amounts B1 to B5, a part of the information is extracted from the information compartment corresponding to the one larger than the above threshold value. On the other hand, all of the information is extracted from the information compartment corresponding to the one smaller than the above threshold value among the information amounts B1 to B5. Here, it is assumed that the magnitude relationship of the information amounts is B3 < threshold value (reference value Bst / 5) < B2. In this case, a part of the information is extracted from the information compartments A1, A2, A4, and A5 having a larger information amount, and the extracted information a1, a2, a4, and a5 are generated. As a part of the information, the information corresponding to the information amount of the above threshold value may be extracted. All of the information is extracted from the information compartment A3 having a smaller information amount, and the extracted information a3 is generated. Assume that they have the information amounts b1, b2, b3, b4, and b5 corresponding to the extracted information a1, a2, a3, a4, and a5 respectively. Here, the magnitude relationship of the information amounts is b1 < B1, b2 < B2, b3 = B3, b4 < B4, b5 < B5. And the total amount of the information amounts b1 to b5 is smaller than the reference value Bst. Note that the "one with a larger information amount" among the first information compartment and the second information compartment corresponds to any one of the information compartments A1, A2, A4, and A5. The "one with a smaller information amount" among the first information compartment and the second information compartment corresponds to the information compartment A3. Also, the threshold value is not limited to the information amount obtained by dividing the reference value Bst into five equal parts as described above, and various modes can be adopted. For example, the following subtraction procedure may be used. First, when it is determined that the total amount of the information amounts B1 to B5 is larger than the reference value Bst, a predetermined value ΔB (< the information amounts B1 to B5, for example, the information amount corresponding to one character or one word may be used) is subtracted from any one of the information amounts B1 to B5. At this time, the information compartments A1 to A5 to be the target of subtracting the predetermined value ΔB may be the ones having the largest amount among the information amounts B1 to B5. Note that the compartment having the largest amount (that is, the target of subtracting the predetermined value ΔB) may be one or more.Next, it is determined again whether the total amount of information B1 to B5 after subtracting a predetermined value ΔB is greater than the reference value Bst. Even after subtracting the predetermined value ΔB, if it is determined that the total amount of information B1 to B5 is greater than the reference value Bst, the predetermined value ΔB is subtracted again from the section with the largest amount of information B1 to B5 after the subtraction, as described above. In this way, the above subtraction process and determination are repeated, and the extracted information at the point in time when it is determined that the total amount of information B1 to B5 after the subtraction is not greater than the reference value Bst corresponds to a1 to a5 (information amounts b1 to b5).

[0029] Referring to Figure 3, the second pattern of information extraction in the information extraction unit 14 will be explained. In the second pattern, the information extraction unit 14 extracts a portion of the information from all information sections A1 to A5, each with a different amount of information, to generate extracted information. This also results in the total amount of information extracted from information sections A1 to A5 being smaller than the reference value Bst.

[0030] More specifically, for example, the total amount of information amounts B1 to B5 in the text information C1 of the information content C may be compared with the reference value Bst of the information amount. Based on this comparison, a part of the information is extracted from all the information sections A1 to A5, respectively. For example, a value less than or equal to the ratio of the total amount of information amounts B1 to B5 to the reference value Bst of the information amount (<1) may be set as a coefficient. Information corresponding to the information amount obtained by multiplying each of the information amounts B1 to B5 by the coefficient is extracted as a part of the information. In this case, a part of the information is extracted from all the information sections A1 to A5, respectively, and the extracted information a1 to a5 is generated. The magnitude relationship of the information amounts is b1 < B1, b2 < B2, b3 < B3, b4 < B4, b5 < B5. And the total amount of the information amounts b1 to b5 is smaller than the reference value Bst. Note that the first information section corresponds to any one selected from the information sections A1 to A5, and the second information section corresponds to the information section other than the arbitrarily selected one. In the second pattern described above, similar to the first pattern, by executing the subtraction procedure described above, the extracted information a1 to a5 may be generated so that the total amount of the information amounts b1 to b5 is smaller than the reference value Bst.

[0031] Referring to FIG. 4, a third pattern of information extraction in the information extraction unit 14 will be described. In the third pattern, the information extraction unit 14 extracts all of the information from a predetermined section and generates the extracted information in the information sections A1 to A5 having different information amounts, respectively, and extracts a part of the information from the other sections and generates the extracted information. Also by this, the total amount of the information extracted from the information sections A1 to A5 becomes smaller than the reference value Bst.

[0032] More specifically, for example, in information content C, the section from which all the text information C1 should be extracted as an information section may be preselected by the user. Here, assuming that the text information C1 in [Claims]·[Means for Solving the Problem] is important, it is assumed that the user has selected information sections A1·A4. In this case, all the information is extracted from the selected information sections A1·A4, and the extracted information a1·a4 is generated. From the unselected information sections A2·A3·A5, a part of the information is extracted, and the extracted information a2·a3·a5 is generated. As a part of the information, information corresponding to the information amount obtained by distributing the information amount less than or equal to the value obtained by subtracting the information amounts B1·B4 from the reference value Bst according to the information amounts B2·B3·B5 may be extracted. Here, the magnitude relationship of the information amounts is b1 = B1, b2 < B2, b3 < B3, b4 = B4, b5 < B5. And the total amount of the information amounts b1~b5 is smaller than the reference value Bst. In the third pattern described above, similar to the first and second patterns, by executing the above-described subtraction procedure, the extracted information a1~a5 may be generated such that the total amount of the information amounts b1~b5 is smaller than the reference value Bst.

[0033] The mode of information extraction by the information extraction unit 14 may be any of the first to third patterns described above. For example, the pattern may be switched according to the characteristics of the information content C or the user's desires. Note that it is not limited to the first to third patterns described above, and other information extraction modes may be adopted.

[0034] As shown in Figure 1, the information processing instruction unit 15 executes an instruction to process the information by combining the multiple extracted pieces of information generated by the information extraction unit 14 so that they become a single piece of information content. As shown in Figures 2, 3, and 4, the extracted pieces of information a1 to a5 generated by extracting from information sections A1 to A5 correspond to text information. The extracted pieces of information a1 to a5 may be combined by taking advantage of the fact that each piece of text has a defined beginning and end. For example, the extracted pieces of information a1, a2, a3, a4, a5 may be combined in the order corresponding to the order of information sections A1, A2, A3, A4, A5. In this case, the end of a1 may be combined with the beginning of a2, the end of a2 with the beginning of a3, the end of a3 with the beginning of a4, and the end of a4 with the beginning of a5. The information obtained by combining the extracted information a1 to a5 in this way contains the characteristics of each of the information sections A1 to A5, while having a smaller amount of information than the reference value Bst, and becomes a new single information content. This new single information content is processed according to the instructions of the information processing instruction unit 15.

[0035] If the information content C includes textual information C1, for example, as in the patent publication of this embodiment, the information processing instruction unit 15 may instruct the processing of text summarization. That is, an instruction may be given to perform text summarization processing on the "information in which the extracted information a1 to a5 described above have been combined (a new single information content)".

[0036] As shown in Figure 1, the information processing instruction unit 15 instructs the server 20 to perform information processing (for example, text summarization) via the information communication network N. The target of information processing in the server 20 is the "information in the state in which the extracted information a1 to a5 described above are combined (a new single information content)". When processing information in the server 20, it is preferable to send the extracted information a1 to a5 generated in the computer 10 from the computer 10 to the server 20. The server 20 may receive the extracted information sent from the computer 10 and store it in the server 20's memory or elsewhere. The information processing unit 21 of the server 20 processes the extracted information a1 to a5 in the state in which they are combined, in accordance with the instructions from the information processing instruction unit 15 of the computer 10. The combination of extracted information a1 to a5 may be performed by either the computer 10 or the server 20. The information processing unit 21 may be configured to be able to perform text summarization as information processing. Here, the text summarization process may be, for example, a process that combines artificial intelligence and natural language processing technology, and automatically generates a summary text from extracted information generated by the computer 10. The instructions of the information processing instruction unit 15 may include, for example, instructions to the information processing unit 21 to generate a summary text, as well as user input such as the desired number of characters / sentences in the summary text and the compression ratio relative to the original information content C. The information such as the summary text generated by the information processing unit 21 may be transmitted from the server 20 to the computer 10 via the information communication network N immediately after generation.

[0037] As shown in Figure 5, the output display unit 16 displays and outputs the summary text generated based on the instructions for processing the document summary. More specifically, the summary text CS generated and transmitted by the server 20 may be displayed on the monitor of the computer 10 or the like. In addition to the summary text, the output display unit 16 also displays and outputs drawing information. More specifically, the drawing information C2 of the information content C acquired from the database 30 may be stored, and the drawing information C2 may be displayed simultaneously with the summary text CS on the monitor of the computer 10 or the like. As for the layout of the output display of the summary text CS and the drawing information C2, for example, multiple pieces of drawing information C2 may be arranged on the left side from the user's perspective, and the summary text CS may be arranged on the right side from the user's perspective. However, this layout is not limited to this, and various forms may be taken. All of the multiple pieces of drawing information included in the information content C may be used for the drawing information C2, or only a portion selected by the user may be used. Furthermore, the drawing information C2 may be automatically resized. For example, the scaling factor of the drawing information C2 may be changed from the original information content C's drawing information depending on the amount of information in the summary text CS. The output drawing information C2 may be selected by the user from multiple drawing information in information content C, or it may be automatically selected according to size, drawing number order, drawing information content, relevance to the user's desired configuration, etc. For example, when automatically selecting drawing information C2 according to the drawing number order, the information processing unit 21 may be instructed to extract only [Figure 1], [Figure 2], and [Figure 3] from the numerous drawing information in information content C. Alternatively, when automatically selecting drawing information C2 based on the relevance to the user's desired configuration, the information processing unit 21 may be instructed to extract the top three drawings with the highest relevance to the user's desired configuration from the numerous drawing information in information content C. Here, "the user's desired configuration" may be, for example, a cross-sectional view of a machine.

[0038] The actual operation of the information processing system 100 will be explained with reference to the flowchart shown in Figure 6 and Figure 1. The information processing system 100 in Figure 1 operates when a program for information processing causes the computer 10 to execute a series of processes in steps S1 to S6 of Figure 6. The information processing program may be stored on a storage medium, which may be installed in either the computer 10 or the server 20, or it may be a portable storage medium separate from the computer 10 and the server 20.

[0039] When the information processing system 100 performs the process of summarizing a document, first, in step S1, the information acquisition unit 12 acquires information content C (in this embodiment, a patent publication including document information C1 and drawing information C2) from the database 30. As a result, as shown by the dashed-dotted arrow [1] in Figure 1, the information content C is transmitted from the database 30 to the computer 10 via the information communication network N. Next, in step S2, the input receiving unit 11 receives the input of the information content C transmitted from the database 30 at the computer 10.

[0040] Next, in step S3, the information partitioning unit 13 partitions the information content C received by the computer 10 so that it contains features, and generates multiple information partitions A1 to A5 containing features. Next, in step S4, the information extraction unit 14 extracts information from the multiple generated information partitions A1 to A5 such that the total amount of information is less than the reference value Bst, and generates multiple extracted information a1 to a5. Any of the first to third patterns described above may be used to generate the extracted information by the information extraction unit 14. The generated extracted information a1 to a5 is transmitted from the computer 10 to the server 20 via the information communication network N, as shown by the solid arrow [2] in Figure 1.

[0041] Next, in step S5, the information processing instruction unit 15 instructs the information processing unit 21 of the server 20 to process the information (processing of text summarization) by combining the multiple extracted information a1 to a5 into a single information content. As a result, the server 20 generates a summary text CS based on the transmitted extracted information a1 to a5. The generated summary text CS is transmitted from the server 20 to the computer 10 via the information communication network N, as shown by the dashed arrow [3] in Figure 1.

[0042] Then, in step S6, the output display unit 16 displays and outputs the summary text CS that has been sent to the computer 10 to the computer 10's monitor, and the series of processes is temporarily terminated. In addition to the summary text CS, drawing information C2 may also be displayed simultaneously on the computer 10's monitor by the output display unit 16 (see Figure 5).

[0043] As described above, the information processing system 100 according to an embodiment of the present invention is an information processing system for processing information contained in an information content C that is recognizable by sight and / or hearing and has a plurality of features. The information processing system 100 according to an embodiment of the present invention includes: an input receiving unit 11 that receives the input of the information content C when the information content C is input; an information partitioning unit 13 that partitions the received information content C so that each of the features is included, and generates a plurality of information partitions A1 to A5 containing the features; an information extraction unit 14 that, when extracting information from the plurality of generated information partitions A1 to A5, extracts information from the information partitions A1 to A5 so that the total amount of information to be extracted is less than a reference value Bst, and generates a plurality of extracted information a1 to a5 corresponding to the plurality of information partitions A1 to A5; and an information processing instruction unit 15 that executes an instruction to process information when the plurality of extracted and generated extracted information a1 to a5 are combined so that they become a single information content.

[0044] According to this, by setting the information quantity reference value Bst to, for example, the upper limit of the amount of information that can be appropriately processed by the information processing unit 21, the information processing unit 21 becomes capable of processing information quantities smaller than the reference value Bst. Therefore, appropriate information processing becomes possible in the information processing unit 21. In particular, the information processing unit 21 often performs information processing using artificial intelligence, and it becomes possible to obtain appropriate output when processing information content using artificial intelligence.

[0045] In the information processing system 100, the information partitioning unit 13 generates at least two first information partitions A1, A2, A4, and A5, each having a different amount of information, and a second information partition A3. The information extraction unit 14 extracts a portion of the information from the first information partition A1, A2, A4, and A5 and the second information partition A3 that has a larger amount of information to generate the extracted information a1, a2, a4, and a5, and extracts all of the information from the partition with a smaller amount of information to generate the extracted information a3 (corresponding to the first pattern; see Figure 2).

[0046] According to this, the amount of information extracted from information sections A1, A2, A4, and A5, which contain a large amount of information, can be kept to a minimum, while all information can be extracted from information section A3, which contains a small amount of information. Therefore, the total amount of information extracted from a1 to a5 can be kept smaller than the baseline value Bst, and information can be extracted from all information sections A1 to A5 without any information being missed, even from information sections with a small amount of information. Such information extraction processing can be easily performed, for example, by setting a threshold for the amount of information to be extracted. As a result, even with information content C that contains a large amount of information, information processing can be easily and evenly performed across all information sections without bias in the scope of information processing.

[0047] In the information processing system 100, the information partitioning unit 13 generates at least a first information partition (one of A1 to A5) and a second information partition (one of A1 to A5 different from the first information partition), each having a different amount of information. The information extraction unit 14 then extracts a portion of the information from both the first and second information partitions to generate the extracted information (corresponding to the second pattern; see Figure 3).

[0048] According to this, a portion of the information can be extracted from each of the information sections A1 to A5. Therefore, the total amount of extracted information a1 to a5 can be kept smaller than the reference value Bst, and information can be extracted from each of the information sections A1 to A5 regardless of the amount of information in each section. This information extraction process can be reliably performed, for example, by setting a coefficient that is less than or equal to the ratio (<1) of the total amount of information B1 to B5 to the reference value Bst. As a result, even with information content C containing a large amount of information, information processing can be reliably and evenly performed across all information sections without bias towards a particular area.

[0049] The information processing system 100 further includes an information acquisition unit 12 that acquires the information content C for input from a database 30 in which a plurality of the information content C are stored.

[0050] According to this, since the database 30 is used to acquire information content C, a large quantity and variety of information content C can be the target of information processing. Furthermore, by searching for information content C in the database 30, the user can easily obtain the information content C they desire. Therefore, when processing information content using artificial intelligence, the target of information processing can be easily made to what the user desires, and a large amount of appropriate output can be obtained.

[0051] In the information processing system 100, the information content C includes at least document information C1, the information partitioning unit 13 generates multiple information partitions A1 to A5 in the document information C1, the information extraction unit 14 generates multiple extracted information a1 to a5 in the document information C1, and the information processing instruction unit 15 instructs the processing of document summarization as the information processing.

[0052] According to this method, even with information content C (for example, a patent publication) containing a large amount of textual information C1, the text summarization process can be performed easily, reliably, and evenly across all information sections, without bias in the scope of information processing. Therefore, an appropriate summarized text CS can be obtained as the output of the information processing.

[0053] The information processing system 100 further includes an output display unit 16 that displays and outputs the summary text CS generated based on the instructions for processing the document summary.

[0054] According to this, for example, the summary text CS can be made to the user via the monitor of computer 10. Therefore, the information content C can be instantly grasped by the user.

[0055] In the information processing system 100, the information content C further includes drawing information C2, and the output display unit 16 displays and outputs the drawing information C2 in addition to the summary text CS.

[0056] According to this, for example, the user can be made aware of drawing information C2 in addition to the summary text CS via the monitor of the computer 10. Therefore, the user can instantly and reliably grasp the information of the information content C. [Explanation of Symbols]

[0057] 10...Computer, 11...Input reception unit, 12...Information acquisition unit, 13...Information partition unit, 14...Information extraction unit, 15...Information processing instruction unit, 16...Output display unit, 20...Server, 21...Information processing unit, 30...Database, A1~A5...Information partition, a1~a5...Extracted information, B1~B5...Information volume, Bst...Reference value, b1~b5...Information volume, C...Information content, C1...Text information, C2...Drawing information, CS...Summary text

Claims

1. In an information processing system for processing information contained in information content that is recognizable by sight and / or hearing and has multiple features, When the aforementioned information content is input, an input receiving unit receives the input of the aforementioned information content, An information partitioning unit divides the received information content so that each of the features is included, and generates a plurality of information partitions containing the features, An information extraction unit that, when extracting information from each of the generated multiple information sections, extracts information from the information sections such that the total amount of information extracted is less than a reference value, and generates multiple extracted pieces of information corresponding to the multiple information sections. An information processing instruction unit executes an instruction to process information, which combines the multiple extracted pieces of information generated above into a single piece of information content, Equipped with Information processing system.

2. In the information processing system described in claim 1, The aforementioned information partition section is, At a minimum, a first information section and a second information section, each having a different amount of information, are generated. The information extraction unit, From the first information section and the second information section, a portion of the information is extracted from the one with the larger amount of information to generate the extracted information, and from the one with the smaller amount of information, all of the information is extracted to generate the extracted information. Information processing system.

3. In the information processing system described in claim 1, The aforementioned information partition section is, At a minimum, a first information section and a second information section, each having a different amount of information, are generated. The information extraction unit, The extracted information is generated by extracting a portion of the information from both the first information section and the second information section. Information processing system.

4. In the information processing system according to claim 2 or claim 3, Information acquisition unit acquires the information content to be input from a database in which multiple information contents are stored. It also has Information processing system.

5. In the information processing system described in claim 4, The aforementioned information content is Includes at least document information, The aforementioned information partition section is, Multiple information sections are generated in the aforementioned text information, The information extraction unit, Multiple pieces of the extracted information are generated from the aforementioned text information, The aforementioned information processing instruction unit is: The aforementioned information processing involves instructing the system to perform text summarization. Information processing system.

6. In the information processing system described in claim 5, An output display unit that displays and outputs the summary text generated based on the instructions for processing the document summary. It also has Information processing system.

7. In the information processing system described in claim 6, The aforementioned information content is Furthermore, including drawing information, The output display unit is, In addition to the summary text, the drawing information is also displayed and output. Information processing system.

8. In an information processing program for processing information contained in information content that is recognizable by sight and / or hearing and has multiple features, When the aforementioned information content is input, an input receiving means for receiving the input of the aforementioned information content is provided. Information partitioning means for partitioning the received information content so that each of the features is included, thereby generating a plurality of information partitions containing the features, Information extraction means for extracting information from each of the generated multiple information sections, such that the total amount of information extracted is less than a reference value, thereby generating multiple extracted pieces of information corresponding to the multiple information sections. Information processing instruction means that executes an instruction to process information, while combining the multiple extracted pieces of information generated above into a single information content, Make the computer execute it. Information processing program.

9. A storage medium storing the information processing program described in claim 8.

Images