Information Processing Apparatus, Information Processing Method, and Program

Abstract

To provide an information processing device capable of checking whether product information on a product is uncertain by mechanical processing, an information processing method, and a program.SOLUTION: An information processing device 30 used in an e-commerce platform comprises one or more processors 31 and one or more memories 32. The memory 32 stores product data 37 on a plurality of products 17 registered in the e-commerce platform. The product data 37 includes a plurality of data sets. Each dataset includes a plurality of data items registered about one product 17. Each data item includes product information on one product 17. The processor 31 is configured to execute indexing processing for one or more product information to obtain an index for comparing one or more product information.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] Patent Document 1 discloses an example of an e-commerce platform where a seller and a purchaser conduct a business transaction. Generally, a seller can register multiple pieces of product information for one product on the e-commerce platform. Examples of product information are the type of product, a description, or an image.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] A purchaser usually examines which product to purchase while comparing similar products based on the registered product information. At this time, the product information registered for a product may be inconsistent, or there may be inconsistencies between the multiple pieces of registered product information. If the registered information is unclear, it can be discovered and improved by carefully checking it with human eyes. However, it is difficult to check all of the huge amount of registered data with human eyes. Therefore, it is desired to establish a method for checking registered information by mechanical processing.

[0005] An object of the present disclosure is to provide an information processing apparatus, an information processing method, and a program that can check whether product information regarding a product is unclear by mechanical processing.

Means for Solving the Problems

[0006] An information processing apparatus according to an aspect of the present disclosure is used in an e-commerce platform. The information processing apparatus includes one or more processors and one or more memories. The memory stores product data related to a plurality of products registered in the e-commerce platform. The product data includes a plurality of data sets. Each data set includes a plurality of data items registered for one of the products. Each data item includes product information related to the one product. The processor is configured to perform an indexing process on the one or more product information in order to obtain an index for comparing the one or more product information.

[0007] An information processing method according to an aspect of the present disclosure is executed by an information processing apparatus used in an e-commerce platform. The information processing method includes: obtaining product data related to a plurality of products registered in the e-commerce platform, where the product data includes a plurality of data sets, each data set includes a plurality of data items registered for one of the products, and each data item includes product information related to the one product; and performing an indexing process on the one or more product information in order to obtain an index for comparing the one or more product information.

[0008] A program according to an aspect of the present disclosure is executed by an information processing apparatus used in an e-commerce platform. The program causes one or more computers to: obtain product data related to a plurality of products registered in the e-commerce platform, where the product data includes a plurality of data sets, each data set includes a plurality of data items registered for one of the products, and each data item includes product information related to the one product; and perform an indexing process on the one or more product information in order to obtain an index for comparing the one or more product information.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

DETAILED DESCRIPTION OF THE INVENTION

[0010] Examples of the information processing apparatus, information processing method, and program of the present disclosure will be described below with reference to the drawings. The present invention is not limited to these examples, but is defined by the claims, and is intended to include all modifications within the meaning and scope equivalent to the claims.

[0011] [INFORMATION PROCESSING SYSTEM] FIG. 1 shows an example of an information processing system 11 related to an e-commerce platform. The information processing system 11 includes a server 20 that provides an e-commerce site for a product 17, and an information processing apparatus 30 used in the e-commerce platform. The server 20 and the information processing apparatus 30 communicate with one or more seller terminals 13 and one or more purchaser terminals 14 through a network 12. The information processing apparatus 30 may be integrated with the server 20. In this case, the server 20 has the functions of the information processing apparatus 30.

[0012] The network 12 includes, for example, the Internet, WAN (Wide Area Network), LAN (Local Area Network), provider terminals, wireless communication networks, wireless base stations, dedicated lines, and the like. It is not necessary for all combinations of the devices shown in FIG. 1 to be able to communicate with each other, and the network 12 may include a partially local network.

[0013] The seller terminal 13 and the purchaser terminal 14 are, for example, smartphones, personal computers, tablets, and the like. Each seller terminal 13 is operated by a seller 15. Each purchaser terminal 14 is operated by a purchaser 16. The seller 15 of the product 17 and the purchaser 16 of the product 17 are users of the business transaction site.

[0014] The server 20 includes one or more processors 21, one or more memories 22, a communication device 23, and a communication bus 24 for connecting these elements to each other. The communication device 23 enables communication with other devices via the network 12, such as the seller terminal 13, the purchaser terminal 14, and the information processing device 30. Stored in the memory 22 are an application 25 for operating the business transaction site and product data 27.

[0015] The server 20 receives product information from the seller terminal 13. Details of the product information will be described later. The received product information is stored in the memory 22 as product data 27 of the product 17 traded on the business transaction site. By storing the product data 27 in this way, the product 17 related to the product data 27 is registered on the business transaction site (electronic commerce platform). It can be said that the product data 27 is registration information regarding the product 17 registered on the business transaction site.

[0016] The purchaser 16 may consider which product 17 to purchase while comparing similar products 17 based on the registered product information. Therefore, the application 25 may have a search function for searching for available products 17 based on various conditions.

[0017] The information processing apparatus 30 includes one or more processors 31, one or more memories 32, a communication device 33, and a communication bus 34 for interconnecting these elements. The communication device 33 enables communication with, for example, the server 20 via the network 12. The memory 32 stores one or more learning programs 35 for machine learning and one or more generated machine learning models 36. The machine learning model 36 can adopt, for example, any one or more of the machine learning models 36a to 36n described later, but is not limited thereto, and any machine learning model that can obtain the target result may be adopted.

[0018] The information processing apparatus 30 periodically, or at a specific timing, or in real time, acquires product data 27 from the server 20. As a result, product data 27 regarding a plurality of products 17 registered in the e-commerce platform is stored as product data 37 in the memory 32. The information processing apparatus 30 may acquire the product data 27 via a component different from the server 20, such as a computer, a server, or a storage. To distinguish it from the product data 27 that is updated at any time in the server 20, the data stored in the information processing apparatus 30 is referred to as product data 37.

[0019] The processors 21 and 31 include arithmetic units such as, for example, CPUs, GPUs, and TPUs. The processors 21 and 31 are processing circuits configured to execute various software processes. The processing circuit may include a dedicated hardware circuit (such as an ASIC, etc.) that processes at least a part of the software process. That is, the software process may be executed by a processing circuit including at least one of one or more software processing circuits and one or more dedicated hardware circuits.

[0020] Memory 22 and memory 32 are computer-readable media. Memory 22 and memory 32 include non-transitory storage media such as, for example, RAM (Random Access Memory), HDD (Hard Disk Drive), flash memory, ROM (Read Only Memory), and the like. Processors 21 and 31 execute a series of instructions included in a program stored in memory 22 and memory 32, respectively, in response to a given signal or in response to a predetermined condition being satisfied.

[0021] [Commerce site] Next, the commerce site provided by server 20 will be described. The commerce site is provided to mediate transactions between seller 15 and purchaser 16. For example, application 25 causes processor 21 to execute a process of storing product information related to a plurality of products 17 received from one or more seller terminals 13 as product data 27 in memory 22.

[0022] As shown in FIG. 2, product data 27 includes a plurality of data sets 50 registered for each product 17. Each data set 50 includes a plurality of data items of different data types. Examples of data items include, but are not limited to, for example, title 51, category 52, description 53, size 54, brand 55, status 56, shipping fee information 57, shipping date 58, region of origin 59, number of "likes" 60, seller profile 61, seller attributes 62, seller evaluations 63, comments to the seller 64, seller registration date 65, seller IP address 66, image 67, price 68. For example, the data item may include audio data instead of or in addition to image 67.

[0023] Each data item includes product information regarding one product 17. Each data set 50 includes at least one product information among a plurality of data items. The types of data included in the plurality of data items may be different from each other. The type of data may be any one of an attribute and an attribute value for the attribute, a character string (text), an image, and a numerical value. Hereinafter, "an attribute and an attribute value for the attribute" may be simply expressed as "attribute: attribute value".

[0024] For example, the types of data of the category 52, size 54, brand 55, shipping fee information 57, shipping time 58, region of the shipping origin 59, and attributes of the seller 62 may be attribute: attribute value. The category 52 and the region of the shipping origin 59 may be set to be selected from predefined categories. The category of the region of the shipping origin 59 is, for example, an administrative district such as a prefecture.

[0025] As shown in FIG. 3, the types of data of the title 51, description 53, status 56, seller's profile 61, and comment 64 to the seller may be character strings. The character string may include a plurality of sentences. The status 56 indicates, for example, a usage status such as "unused", "new", and "no noticeable scratches or dirt". The type of data of the image 67 is an image, and the image may include a video. The image 67 may be a still image including characters (in FIG. 3, "the product is a photo"), or may be a video including audio data. The types of data of the "like" number 60, seller's evaluation 63, and price 68 may be numerical values.

[0026] The title 51 may include one or more pieces of information indicating target attributes (for example, name, brand name, size, and color, etc.). The title 51 may be, for example, a long character string including all of the name, brand name, size, and color. Further, the title 51 may include a promotional phrase for sales promotion, such as "limited", "free shipping", "recommended", which is not an attribute of the product 17 itself.

[0027] The title 51 and the description 53 may have character count limitations, and the maximum number of characters for the title 51 may be less than that of the description 53. The title 51 is also the identification information of the product 17. Therefore, as a data item, identification information (for example, the name of the product 17, etc.) may be included instead of the "title 51".

[0028] The application 25 causes the processor 21 to execute a process of displaying the registered product data 37 on the product screen 18 shown in FIG. 3. The product screen 18 has a display area for displaying each data item. The product screen 18 is displayed on the display of each user terminal (for example, the seller terminal 13 or the purchaser terminal 14) in response to a request from a user (for example, the seller 15 or the purchaser 16).

[0029] [Normalization processing] The processor 31 is configured to perform a normalization process on one or more pieces of product information in order to obtain an index 71 for comparing one or more pieces of product information. At least a part of the normalization process may be executed using one or more machine learning models 36.

[0030] Here, when obtaining two indicators 71 based on the product information included in any two of the plurality of data items, those two data items may be referred to as the first data item and the second data item. That is, the first, second, third, etc. do not refer to a specific object (such as a data item, product information, or indicator), but are ordinal numbers used to identify a plurality of data items. The number of ordinal numbers is not limited to two and can be changed to any number of three or more. In this case, the product information included in the first data item may be referred to as the first product information, and the product information included in the second data item may be referred to as the second product information. Also, the indicator 71 obtained based on the first product information may be referred to as the first indicator 71a, and the indicator 71 obtained based on the second product information may be referred to as the second indicator 71b. In this case, the indexing process includes obtaining a first indicator 71a related to one product 17 based on the first product information and obtaining a second indicator 71b related to the same product 17 based on the second product information. Then, the processor 31 is configured to execute a comparison process for comparing a plurality of indicators 71 (for example, the first indicator 71a and the second indicator 71b).

[0031] The indicator 71 is, for example, a category defined for classifying the product 17. In this case, obtaining the indicator 71 includes classifying one product 17 based on the product information. The product 17 may be classified so as to correspond to any one of the categories 52 which are one of the data items. Alternatively, the product 17 may be classified in a category different from the category 52 which is a data item.

[0032] Product 17 may be classified using a known machine learning model 36a based on product information. For example, by inputting a character string into a convolutional neural network (CNN) or BERT (Bidirectional Encoder Representations from Transformers) for natural language processing (NLP) to solve a classification task, a classification result (index 71) may be obtained. Similarly, by inputting an image into a machine learning model 36b such as a CNN for image processing to solve a decomposition task, a classification result (index 71) can be obtained.

[0033] The indexing process may include obtaining two or more classification results by classifying one product 17 two or more times based on different information. For example, as a first embodiment, when category 52 includes "doll" and "photo", classifying product 17 as "doll" based on description text 53 which is the first data item results in obtaining the first index 71a, and classifying product 17 as "photo" based on image 67 which is the second data item results in obtaining the second index 71b.

[0034] After executing a comparison process for comparing two or more such classification results, the processor 31 may be configured to execute a specifying process based on the comparison result. The specifying process may specify that the product information (at least the product information used for classification) registered for one product 17 is unclear when two or more classification results do not match each other, and may include specifying the category of that one product 17 based on the classification result when two or more classification results match each other. Regarding the match of the comparison result, it does not necessarily have to be an exact match, and depending on the type of index 71, there may be cases where it suffices to match within a specific condition (for example, a threshold condition).

[0035] As a first embodiment, when the first index 71a (the first classification result) is "doll" and the second index 71b (the second classification result) is "photo", it is specified that it is unclear whether the product 17 is a doll or a photo. Such ambiguity can occur when there is a misrecording in the registered product information or when outdated information is registered.

[0036] On the other hand, if all of the plurality of indices 71 have the same classification result (for example, "doll"), those classification results are likely to be correct. Therefore, based on the classification results, the category of the product 17 may be specified as "doll".

[0037] In a simpler second embodiment, the category 52 which is the product information itself may be used as the first index 71a, and this first index 71a may be compared with the second index 71b which is the classification result obtained from other product information. In this case, the processor 31 may execute an indexing process on one product information in order to compare two pieces of product information. And the processor 31 may be configured to execute a specifying process based on two or more indices 71 (for example, the first index 71a and the second index 71b). This specifying process includes specifying that the product information registered for one product 17 is unclear when two or more indices 71 do not match each other, and specifying the category 52 of the one product 17 based on the matching indices 71 when two or more indices 71 match each other.

[0038] It can be said that the clarity of the registration information is higher as the number of matching classification results is larger. Therefore, a specifying process may be performed based on three or more indices 71 (three or more classification results). In this case, it may be specified as clear when all of the indices 71 match, or there may be a case where it is sufficient if they match within a specific condition range. For example, when it can be determined that the obtained index 71 is invalid or inappropriate, a specifying process may be performed based on the other indices 71 excluding that index 71.

[0039] The index 71 may be the probability that the product 17 belongs to a certain category 52. In this case, the indexing process may include calculating the probability that one product 17 belongs to a certain category 52. Then, when the probability that a product belongs to a certain category 52 is equal to or greater than a pre-specified threshold value, the processor 31 may determine that the product 17 belongs to that category 52.

[0040] The indexing process may include generating processed information by processing product information prior to classification. In this case, obtaining the index 71 for one product based on the product information includes obtaining the index 71 based on the processed information. For example, the indexing process may include a process of classifying the product 17 based on the processed information.

[0041] The indexing process may include obtaining a first index 71a related to one product 17 based on one product information, generating processed information by processing the one product information, and obtaining a second index 71b related to the one product 17 based on the processed information. For example, as a third embodiment, the first index 71a is obtained by classifying the product 17 as a "doll" based on the description text 53 (original text). Further, a processed text obtained by processing the same description text 53 is generated as processed information. Then, the second index 71b is obtained by classifying the same product 17 as a "doll" based on the processed text. In this case, the processor 31 is configured to execute a comparison process and a specification process for comparing the first index 71a and the second index 71b.

[0042] If the first index 71a is "doll" and the second index 71b is also "doll", the product 17 is specified as a doll. On the other hand, if two or more indexes 71 are different from each other, it is specified that the registered product information is unclear. Such processing is particularly effective when the description text 53 is long, when there is no image 67 or the image 67 is unclear, or when the number of registered data items is small. For example, when there is no image 67, it is necessary to determine the content from other product information such as the description text 53. However, since the purchaser 16 may skip or misread the content when the description text 53 is long.

[0043] Processing the product information may include any one or more of encoding the product information, summarizing the character string that is the product information, processing the image that is the product information, extracting a part from the product information, and converting voice data into a character string (for example, "This product is not a stuffed toy"). For example, in the third embodiment above, a summary sentence obtained by summarizing the description 53 may be used as the processed information.

[0044] The summarization of the character string can be performed, for example, using a natural language processing model. The character string may be a character string converted from voice data or a character string included in an image. The summary may be an extractive summary, an abstractive summary, or a combination thereof.

[0045] The extractive summary model is configured to generate a summary sentence by extracting important parts from the text. For example, the description 53, for example, is input to a machine learning model 36c that has been pre-trained on the importance according to the distributed (vector) representation obtained from the character string. As an example, using a convolutional neural network, the input description 53 is divided into a plurality of parts (for example, a plurality of sentences), and a classification problem of whether each divided part is important or not is solved. Thereby, the distributed representation of each part is obtained. Then, a summary sentence is generated by arranging the parts classified as important. The summary sentence thus created can be said to be a patched-together sentence of parts extracted from the original description 53, or a sentence obtained by partially deleting the original description 53.

[0046] Alternatively, the importance of the distributed representation of each part obtained by a convolutional neural network may be evaluated by another model such as a sequence-to-sequence model. In this case, a summary sentence is generated by arranging the parts evaluated as important. In addition, an extractive summary model using a pre-trained machine learning model 36d such as BERT or ELECTRA can also be used to generate a summary sentence.

[0047] The abstractive summarization model is configured to generate a new text (summary text) from the original text, such as the explanatory text 53. For example, a sequence-to-sequence model may be used. Alternatively, an abstractive summarization model using a pre-trained machine learning model 36e such as BERT or ELECTRA may be used. Furthermore, a machine learning model 36f of a type that mixes extractive summarization and abstractive summarization may be used.

[0048] To train or fine-tune such a model, the data may be labeled. For example, if a characteristic word is included in each category (for example, if the word "photo" is included in a text where the category is "card"), that text may be used as correct answer data. Thereby, a machine learning model 36g for summarization can be trained or fine-tuned according to the category. Alternatively, characteristic words or terms that are likely to become unclear for a certain category may be registered in a database, and the registered terms may be included in the summary text.

[0049] The image processing may be super-resolution processing for increasing the resolution of the image or processing for decreasing the resolution of the image. Instead of this, or in addition to this, image processing such as optimization of contrast, brightness, exposure, and other parameters may be performed. Alternatively, a characteristic part (for example, a human figure part) may be cut out (cropping) from the image 67, or the background part of the human figure may be cut off (trimming). The cutting out or cutting off of the image can also be said to be information extraction. The change in resolution or parameters may be performed on the extracted image (hereinafter referred to as "extracted image").

[0050] For example, when generating an extracted image (processed information) from the original image 67, characteristic parts may be extracted using any map model such as an activation map or a saliency map. The activation map can be generated, for example, by applying Grad-CAM (Gradient-weighted Class Activation Mapping) to the image input to the convolutional neural network model. Then, the activated regions (regions indicating a certain category) are cut out, or the non-activated regions (regions not indicating a certain category) are cut out. Alternatively, using the saliency map, processed information may be generated by cutting out regions with low saliency scores (regions that are easily overlooked by people).

[0051] As another method, image cropping may be performed based on the map of the aesthetic score. For example, Grad-CAM is applied to a CNN-based image evaluation model such as NIMA to generate a heat map that emphasizes the regions of the image according to the aesthetic score. Thereby, in the image, processed information may be generated by cutting out regions with high aesthetic scores or cutting out regions with low aesthetic scores (regions that are easily overlooked by people).

[0052] Symbolization includes converting a plurality of different types of data (e.g., category 52, description text 53, and image 67) into comparable data. Symbolization may be performed on the processed information. As an example of symbolization, product information, such as a character string and an image, may be converted into a distributed representation respectively. For example, in the above first embodiment, the description text 53 (original text) and the image 67 may be converted into distributed representations respectively. For example, a known machine learning model 36h such as CNN, FastText, Doc2Vec, Sentence2Vec, Data2Vec, or BERT may be used to symbolize the character string. When this machine learning model 36h receives a character string as input, it outputs a distributed representation of that character string. The symbolized data, such as the distributed representation, is an example of the index 71. The symbolization of the image can be performed using a known machine learning model 36i such as CNN.

[0053] Extracting a part from the product information may include at least one of extracting a part from the character string that is the product information, extracting characters from the image that is the product information, cutting out or cropping a part of the image that is the product information, and extracting audio data or a character string from a video.

[0054] When extracting characters from an image, first, the characters are detected from the image, and then the detected characters are recognized to form a character string. For this process, a machine learning model 36j that performs character detection and character recognition as separate tasks may be adopted, or a machine learning model 36k that performs character detection and character recognition simultaneously may be adopted. The extraction of characters can be performed using a machine learning model 36m for known OCR (Optical Character Recognition). For example, one-shot object detection such as CNN, Faster R-CNN, R2CNN (Rotational Region CNN), YOLO (You Only Look Once), FOTS (Fast Oriented Text Spotting), etc. may be adopted. Alternatively, a known OCR application may be used to extract a character string from the image.

[0055] When extracting a character string from video or audio data, automatic speech recognition (ASR) that recognizes speech and converts it into a character string is performed. The automatic speech recognition can be performed using, for example, a machine learning model 36n that applies a recurrent neural network (RNN), a Residual CNN, a ContextNet, or a Transformer.

[0056] The extracted character string or image may be encoded or classified as necessary and compared with the original product information, other product information, or other processed information. For example, the result of classifying the entire image 67 may be compared with the result of classifying the extracted image of the same image 67. When a plurality of different images 67 are registered, the above-described indexing process may be performed for each image 67.

[0057] As an example, the characters extracted from the image 67 (in FIG. 3, "The product is a photo") can be used as processed information. As a fourth embodiment, the classification result obtained by encoding and then classifying the extracted characters may be compared with the classification result based on other product information, for example, the description text 53 or the processed information of the description text 53 (for example, the encoded summary text). Alternatively, the characters extracted from the characteristic (important) expressions (in FIG. 3, for example, "Photo of a plush toy set") in the text of the description text 53 may be encoded.

[0058] When two distributed representations (two pieces of product information, two indicators 71, or two classification results) have a similarity equal to or higher than a preset threshold, the two distributed representations may be specified as matching. In other words, if the similarity of the two distributed representations is less than the threshold, it can be specified that the product information registered for one product is unclear.

[0059] When three or more distributed representations are obtained, when there are two dissimilar distributed representations, the processor 31 may specify that the product information is unclear. When a more rigorous judgment is required, if there is even one distributed representation with a similarity less than the threshold, the processor 31 may specify that the product information is unclear.

[0060] In the fifth embodiment shown in FIG. 4, the first index 71a is "probability that product 17 is a photo: 100%" based on the processed information obtained by encoding the characters extracted from the image 67. The second index 71b is "probability that product 17 is a photo: 50% and probability that the same product 17 is a doll: 50%" based on the processed information obtained by encoding the description text 53. The third index 71c is "probability that product 17 is a photo: 100%" based on the processed information obtained by encoding the summary text of the description text 53. When such three indexes 71 (the first index 71a to the third index 71c) are obtained, the processor 31 may specify that the product information is unclear based on the fact that the second index 71b includes "probability of being a doll: 50%".

[0061] In the specifying process, simply a binary specifying result of "clear" or "unclear" may be set. In addition to this, a plurality of intermediate specifying results such as "generally clear" or "possibly unclear" may be set according to the probability or the degree of variation of the index 71. Alternatively, as the specifying result, a frequency indicating clarity (for example, clarity: 1, 2, 3 ···) may be adopted.

[0062] As in the sixth embodiment shown in FIG. 5, one piece of product information, for example, the description text 53 may be divided into a plurality of parts 53a to 53e, for example, one sentence at a time, and each part (each sentence) may be encoded to obtain probabilities for each part (for each sentence), for example, the first index 71a to the fifth index 71e. Then, when the probability of a category different from the other indexes 71 among the first index 71a to the fifth index 71e obtained from one piece of product information is equal to or higher than a threshold value (for example, 50% or 100%), the processor 31 may specify that the product information is unclear.

[0063] As in the seventh embodiment shown in FIG. 6, using tokenization, one piece of product information, for example, the description text 53, may be decomposed (tokenized) into words (tokens), and each of the decomposed words may be encoded in a computer-processable format. Then, the probability (indices 71a, 71b, 71c...) corresponding to a specific category may be calculated for each word.

[0064] That is, the unit for splitting the character string may be a "sentence" or a "word". Alternatively, the character string may be split into arbitrary units such as for each paragraph, each section, or each specific number of characters. And among the plurality of indices 71 obtained for each of those parts, when the probability that it belongs to a category different from the other indices 71 is equal to or greater than a threshold value, the processor 31 may specify that the product information is unclear. Further, in addition to the probability for each part, the probability of the entire description text 53 may be used as an additional index 71.

[0065] The method of judgment or the criteria for judgment when performing specific processing based on specific product information may be changed. For example, when the price 68 exceeds a specified threshold value (for example, a price that greatly exceeds the average price in a certain category 52 or attribute), the number of indices 71 to be compared may be increased, or the criteria for specifying as clear may be made stricter. Alternatively, when there is a category 52 in which the product information tends to be unclear, for the product 17 in the corresponding category 52, the number of indices 71 to be compared may be increased, or the criteria for specifying as clear may be made stricter. In addition, the method of judgment or the criteria for judgment may be changed according to the attribute 62 of the seller.

[0066] When a certain condition is satisfied, an index 71 based on specific product information may be obtained and used as a comparison target. For example, when the number of characters in the description text 53 is equal to or greater than a specified number, an index 71 based on the processed information of the description text 53 may be obtained. Alternatively, when the evaluation 63 of the seller 15 is lower than a specified level, for the product 17 of the seller 15, the number of indices 71 to be compared may be increased, or the criteria for specifying as clear may be made stricter.

[0067] [Information Processing Executed by the Information Processing Apparatus 30] FIG. 7 shows an example of information processing executed by the information processing apparatus 30 to identify whether the registered product information is unclear. For each of a plurality of products 17 registered in the e-commerce platform, the information processing apparatus 30 executes the processes of steps S11 to S15. This information processing may be performed each time a product 17 is registered, or may be continuously performed for a plurality of products 17 for each category 52 or at a specified timing.

[0068] In step S11, the processor 31 acquires product data 27 from the server 20 and stores it in the memory 32 as product data 37. In step S12, the processor 31 performs an indexing process on one or more pieces of product information in order to obtain an index 71 for comparing one or more pieces of product information for one product 17. The product information to be subjected to the indexing process in step S12, the content of the indexing process, and the number of indexes 71 to be obtained may be changed according to conditions such as the product 17, the category 52, or the registered product information.

[0069] In step S13, the processor 31 executes a comparison process for comparing at least two of the one or more indexes 71 obtained in step S12. More specifically, the processor 31 determines whether a plurality of comparison targets match.

[0070] In step S14, the information processing apparatus 30 executes a specifying process based on the comparison result of step S13. For example, when the plurality of indexes 71 do not match or when the requirements related to the index 71 are not satisfied (for example, when the probability of being a certain category is less than the threshold), the processor 31 specifies that the product information registered for the product 17 is unclear. On the other hand, when the plurality of indexes 71 match or when the requirements related to the index 71 are satisfied, the processor 31 specifies that the product information registered for the product 17 is clear. When the index 71 is a category, instead of or in addition to the product information being clear, the processor 31 may specify that the product 17 belongs to the category obtained as the index 71.

[0071] In step S15, the processor 31 stores at least the result of the specific process in the memory 32 and ends the process. In step S15, the processor 31 may further store the index 71 obtained in step S12 and the comparison result of step S13 in the memory 32.

[0072] [Operation of the present disclosure] When a plurality of pieces of product information are registered for one product 17, it is not easy to determine whether all of those pieces of product information are correct. For example, it is difficult for a third party (for example, the administrator of an e-commerce site) to confirm whether the registered image 67 is a photo of the registered product 17. Similarly, when there is an inconsistency between a plurality of pieces of product information, it is difficult to determine which piece of product information is incorrect. There are various patterns of deficiencies related to product information, and it takes a great deal of effort to define all of the judgment criteria. In addition, when the data types of a plurality of pieces of product information are different from each other, they cannot be directly compared.

[0073] In that regard, the information processing apparatus 30 can obtain a plurality of indexes 71 by respectively indexing a plurality of pieces of product information. In this case, even if the data types of the product information are different, comparable indexes 71 can be obtained by performing processes such as encoding or classification. Furthermore, the information processing apparatus 30 can obtain indexes 71 from the original product information and the processed information respectively by processing one piece of product information.

[0074] Then, by comparing the plurality of indexes 71 by the information processing apparatus 30, at least it is possible to check whether the product information is unclear. That is, even if it is not possible to specify what kind of deficiency exists in which piece of product information, it is possible to specify or estimate whether the product information is unclear by mechanical processing.

[0075] Since the normalization process can be performed by data processing using the machine learning model 36, the labor of individually checking with human eyes can be saved. In order to perform the normalization process, the machine learning model 36 corresponding to the type of data can be selected, so that it is possible to handle various product information.

[0076] By checking whether the product information is unclear or not, it becomes possible to take various measures to correct the unclear information. For example, if only the product information identified as unclear is checked by human eyes, it is possible to save a significant amount of labor compared to checking all product information by human eyes. Also, if the patterns related to the lack of information can be grasped through such checks, it becomes possible to establish separate judgment methods for each pattern. By such measures, the reliability and clarity of the registration information on the commercial transaction site can be enhanced.

[0077] [Effects of the Present Disclosure] According to the present disclosure, the following effects can be achieved. (1) Using the information processing apparatus 30, it is possible to check the uncertainty of the product information regarding the product 17 by mechanical processing.

[0078] (2) As in the third embodiment, by comparing the first index 71a obtained from one original product information with the second index 71b obtained from the processed information of the product information, the uncertainty of the product information can be checked. Therefore, comparison processing can be executed using a plurality of indices 71 obtained from one product information.

[0079] (3) As in the first embodiment, different indices 71 can be obtained from a plurality of product information. And by comparing those indices 71, the uncertainty of the product information can be checked.

[0080] (4) By processing the product information, it becomes possible to obtain more diverse information different from the original product information. Also, different indices 71 can be obtained from the original product information and its processed information, respectively.

[0081] (5) By encoding product information, it becomes possible to compare different types of data with each other. (6) By summarizing a character string, the original product information (e.g., description text 53) can be converted into a more characteristic expression that is more eye-catching when read by a user (e.g., purchaser 16).

[0082] (7) By processing the image 67, it can be converted into a more characteristic image that is more eye-catching when seen by a user (e.g., purchaser 16). (8) By extracting a part from product information such as an image or a character string, a more characteristic part of the product information can be extracted. This enables a more accurate check to be performed.

[0083] (9) By extracting a part from a character string, an index 71 that emphasizes a more characteristic part (e.g., a word) can be obtained. (10) By extracting characters from the image 67, an index 71 can be obtained based on the character information contained in the image 67.

[0084] (11) By cutting out or cropping a part of the image 67, an index 71 can be obtained based on a more characteristic part. (12) By using the category of the product 17 as the index 71, it becomes possible to compare different types of product information with each other. Also, by changing the fineness of classification, it becomes possible to adjust the computational amount of the indexing process.

[0085] (13) By using the probability that the product 17 belongs to a certain category as the index 71, it becomes possible to quantitatively determine whether the product information is unclear by comparing the probability with a threshold value.

[0086] (14) Although the plurality of data items contain different types of data from each other, by performing an indexing process including encoding, it becomes possible to compare those data. By using a machine learning model 36 of 1 or more, it becomes possible to perform an indexing process close to human judgment. Also, by combining a plurality of machine learning models 36, it becomes possible to perform an indexing process for different types of data.

[0087] This embodiment can be implemented with the following modifications. This embodiment and the following modification examples can be implemented in combination with each other within a technically non - conflicting range. · The index 71 may be the probability of being a certain object (for example, a brand, item, type, attribute, attribute value, or a combination thereof related to a certain product 17, or a character used as a motif, etc.).

[0088] · The index 71 may be a similarity. For example, as an indexing process, a plurality of product information may be vectorized respectively, and it may be determined to which category it belongs based on the cosine similarity between each other. Alternatively, based on the similarity to a value (correct data) obtained by indexing a certain object (for example, a brand, item, type, attribute, attribute value, or a combination thereof related to a certain product 17, or a character used as a motif, etc.), the category of the indexed product information may be determined.

[0089] · The category 52 may be a data item generated based on the classification result specified by the information processing apparatus 30, rather than the product information registered by the seller 15. Alternatively, when the category 52 is not registered by the seller 15, data of the category 52 may be generated based on the classification result by the information processing apparatus 30. As another example, when the registered product information is unclear, the category 52 registered by the seller 15 may be corrected based on the specified classification result.

[0090] The aspects grasped from the above embodiment and modification examples are listed below. [1] An information processing apparatus used in an e - commerce platform, The information processing apparatus includes 1 or more processors and 1 or more memories, The memory stores product data regarding a plurality of products registered in the e-commerce platform, the product data includes a plurality of data sets, each of the data sets includes a plurality of data items registered for one of the products, each of the data items includes product information regarding the one product, the processor is configured to perform an indexing process on the one or more product information to obtain an index for comparing the one or more product information. Information processing apparatus.

[0091] According to this configuration, a plurality of indexes can be obtained from one product information, or an index can be obtained from each of a plurality of product information. In this way, by obtaining a plurality of indexes from the product information, it becomes possible to compare the product information.

[0092] [2] The indexing process includes obtaining a first index related to the one product based on the one product information, generating processed information by processing the one product information, obtaining a second index related to the one product based on the processed information, and the processor is further configured to perform a comparison process of comparing the first index and the second index. The information processing apparatus according to [1] above.

[0093] According to this configuration, an index can be obtained from one product information and the processed information of the product information, respectively. Therefore, it becomes possible to obtain a plurality of indexes from one product information. In addition, the processor may further perform a specific process based on the comparison result in the comparison process after the comparison process. In this specific process, when the first index and the second index do not match, it may be specified that the product information related to the indexing is unclear.

[0094] [3] The plurality of data items includes a first data item and a second data item, the first data item includes first product information which is the product information, and the second data item includes second product information which is the product information, The indexing process is obtaining a first index related to the one product based on the first product information, obtaining a second index related to the one product based on the second product information, and the processor is further configured to execute a comparison process of comparing the first index and the second index. The information processing apparatus according to [1] above.

[0095] According to this configuration, an index can be obtained from each of a plurality of product information. Therefore, it becomes possible to mechanically compare a plurality of different product information via the index obtained by the indexing process. Note that the processor may further execute a specific process based on the comparison result in the comparison process after the comparison process. In this specific process, when the first index and the second index do not match, it may be specified that the product information related to the indexing is unclear.

[0096] [4] The indexing process includes generating processed information by processing the product information, obtaining the index related to the one product based on the product information includes obtaining the index based on the processed information. The information processing apparatus according to [3] above.

[0097] According to this configuration, an index can be obtained from one product information and the processed information of the product information. Therefore, it becomes possible to obtain a plurality of indexes from one product information.

[0098] [5] Processing the product information includes any one or more of encoding the product information, summarizing a character string which is the product information, processing an image which is the product information, and extracting a part from the product information. The information processing apparatus according to any one of [2] to [4] above.

[0099] According to this configuration, various processed information can be obtained by encoding, summarizing, image processing, or extraction. Then, by performing appropriate processing according to the original product information, it becomes possible to obtain an index that more clearly shows the characteristics than the original product information.

[0100] [6] The extraction includes at least one of extracting a part from the character string that is the product information, extracting characters from the image that is the product information, and cutting out or cropping a part of the image that is the product information. The information processing apparatus according to claim 5.

[0101] According to this configuration, by extracting characteristic parts from the character string, it becomes possible to obtain an index that more clearly shows the characteristics than the original product information. Also, by extracting characters from the image, it becomes possible to obtain data of a different type from the original product information. Also, by cutting out a characteristic area from the image, it becomes possible to obtain an index that more clearly shows the characteristics than the original product information. Also, by cropping an area without characteristics such as the background from the image, a more characteristic area can be left. Thereby, it becomes possible to obtain an index that more clearly shows the characteristics than the original product information.

[0102] [7] The index is a category defined for classifying the product. Obtaining the index includes classifying the one product based on the product information. The information processing apparatus according to any one of [1] to [6] above.

[0103] According to this configuration, by classifying various product information into respective defined categories, it is possible to make indices that can be compared with each other by mechanical processing. [8] The index is the probability that the product belongs to a certain category. The normalization process includes calculating the probability that the one product belongs to the certain category. The information processing apparatus according to any one of [1] to [6] above.

[0104] According to this configuration, for each of various product information or processed information, by calculating the probability that one product belongs to a certain category, the probabilities can be compared with each other. Thereby, it becomes possible to determine to what extent a plurality of indicators match by mechanical processing.

[0105] [9] The plurality of data items include data of different types from each other. The type of the data is any one of an attribute and an attribute value for the attribute, a character string, an image, and a numerical value. The information processing apparatus according to any one of [1] to [8] above.

[0106] The information processing apparatus according to claim 1. According to this configuration, by normalizing a plurality of data of different types, it becomes possible to compare a plurality of product information with each other by mechanical processing.

[0107]

[10] The plurality of data items include any one or more of a title, a size, a brand, a status, a description, an image, and a category regarding the product. The information processing apparatus according to any one of [1] to [9] above.

[0108] According to this configuration, even if the data items registered for each product are not unified, by normalizing any one or more of the data items, it becomes possible to mechanically check the uncertainty of the product information about the product.

[0109]

[11] The plurality of data items include any one or more of a profile of the seller of the product, an evaluation of the seller, a comment on the seller, a registration date of the seller, and an IP address of the seller. The information processing apparatus according to any one of [1] to

[10] above.

[0110] According to this configuration, it becomes possible to mechanically check the ambiguity of product information based on data items related to sellers who sell each product. As a result, it is possible to obtain credit information regarding each seller.

[0111]

[12] At least a part of the indexing process is executed using one or more machine learning models. The information processing apparatus according to any one of [1] to

[11] above.

[0112] According to this configuration, by using one or more machine learning models, an indexing process close to human judgment becomes possible. Also, by combining a plurality of machine learning models, an indexing process for different types of data becomes possible.

[0113]

[13] The index is a category defined for classifying the product. The indexing process includes obtaining two or more classification results by classifying the one product two or more times based on different information. The processor is configured to execute a specific process based on the two or more classification results, and the specific process is When the two or more classification results do not match each other, it is specified that the product information registered for the one product is unclear, and when the two or more classification results match each other, the category of the one product is specified based on the classification result. The information processing apparatus according to any one of [1] to

[12] above.

[0114] According to this configuration, based on two or more classification results, it becomes possible to specify whether the product information registered for one product is unclear. Also, when the two or more classification results match each other, it becomes possible to specify the category of that product.

[0115]

[14] The index is a category defined for classifying the product. The plurality of data items includes the category of the product, The indexing process includes obtaining a classification result by classifying the one product based on the product information, The processor is configured to execute a specifying process based on the category and the classification result, and the specifying process identifies that the product information registered for the one product is unclear when the category and the classification result do not match each other, and includes specifying the category of the one product based on the classification result when the category and the classification result match each other. The information processing apparatus according to any one of [1] to

[12] above.

[0116] According to this configuration, it becomes possible to identify whether the product information is unclear based on the category which is the product information and the classification result of product information different from the category.

[15] An information processing method executed by an information processing apparatus used in an e-commerce platform, obtaining product data regarding a plurality of products registered in the e-commerce platform, the product data including a plurality of data sets, each data set including a plurality of data items registered for one of the products, and each data item including product information regarding the one product, executing an indexing process on the one or more product information to obtain an index for comparing the one or more product information, An information processing method including.

[0117]

[16] A program executed by an information processing apparatus used in an e-commerce platform, The program causes one or more computers to Obtaining product data regarding a plurality of products registered on the e-commerce platform, wherein the product data includes a plurality of data sets, each data set includes a plurality of data items registered for one of the products, and each data item includes product information regarding the one product, Performing an indexing process on the one or more product information to obtain an index for comparing the one or more product information, A program for causing the execution.

Explanation of Signs

[0118] 11…System, 12…Network, 13…Seller Terminal, 14…Purchaser Terminal, 15…Seller, 16…Purchaser, 17…Product, 18…Product Screen, 20…Server, 21…Processor, 22…Memory, 23…Communicator, 24…Communication Bus, 25…Application, 27…Product Data, 30…Information Processing Device, 31…Processor, 32…Memory, 33…Communicator, 34…Communication Bus, 35…Learning Program, 36, 36a, 36b, 36c, 36d, 36e, 36f, 36g, 36h, 36i, 36j, 36k, 36m, 36n…Machine Learning Model, 37…Product Data, 50…Data Set, 51…Title, 52…Category, 53…Description, 67…Image, 68…Price, 71…Index, 71a…First Index, 71b…Second Index, 71c…Third Index.

Claims

1. An information processing apparatus used in an e-commerce trading platform, wherein the information processing apparatus includes a memory configured to store product information corresponding to each of a plurality of data items of products registered in the e-commerce trading platform, obtaining a first indicator indicating a category of the product based on the product information corresponding to a first data item among the plurality of data items, obtaining a second indicator indicating the category of the product based on the product information corresponding to a second data item among the plurality of data items, comparing the first indicator and the second indicator, and when the first indicator and the second indicator do not match, identifying that the product information is unclear, and is configured to execute, an information processing apparatus.

2. An information processing apparatus used in an e-commerce trading platform, wherein the information processing apparatus includes a memory configured to store product information corresponding to each of a plurality of data items of products registered in the e-commerce trading platform, obtaining a first indicator indicating a category of the product based on the product information corresponding to a first data item among the plurality of data items, obtaining a second indicator indicating the category of the product based on the product information corresponding to a second data item among the plurality of data items, when the similarity between the first indicator and the second indicator is less than a threshold value, identifying that the product information is unclear, and is configured to execute, an information processing apparatus.

3. An information processing apparatus used in an e-commerce trading platform, wherein the information processing apparatus includes a memory configured to store product information corresponding to each of a plurality of data items of products registered in the e-commerce trading platform, obtaining a first indicator indicating a probability that the product belongs to a certain category based on the product information corresponding to a first data item among the plurality of data items, obtaining a second indicator indicating a probability that the product belongs to the certain category based on the product information corresponding to a second data item among the plurality of data items, identifying that the product information is unclear based on the first indicator and the second indicator, and is configured to execute, an information processing apparatus.

4. generating processed information by processing product information corresponding to at least one of the plurality of data items, obtaining at least one of the first index and the second index based on the processing information The information processing apparatus according to any one of claims 1 to 3.

5. The processing of the product information includes any one or more of encoding the product information, summarizing the character string that is the product information, processing the image that is the product information, and extracting a part from the product information. The information processing apparatus according to claim 4.

6. The extracting includes at least one of extracting a part from the character string that is the product information, extracting characters from the image that is the product information, and cutting out or cropping a part of the image that is the product information. The information processing apparatus according to claim 5.

7. The plurality of data items include data of different types from each other. The type of the data is any one of an attribute and an attribute value for the attribute, a character string, an image, and a numerical value. The information processing apparatus according to any one of claims 1 to 3.

8. The plurality of data items include any one or more of a title, a size, a brand, a status, a description, an image, and a category related to the product. The information processing apparatus according to any one of claims 1 to 3.

9. The plurality of data items include any one or more of a profile of the seller of the product, an evaluation of the seller, a comment on the seller, a registration date of the seller, and an IP address of the seller. The information processing apparatus according to any one of claims 1 to 3.

10. At least one of the first index and the second index is obtained using one or more machine learning models. The information processing apparatus according to any one of claims 1 to 3.

11. Further including, when the product information is not ambiguous, specifying the category indicated by the first index and the second index as the category of the product. The information processing apparatus according to any one of claims 1 to 3.

12. An information processing method executed by an information processing apparatus used in an e-commerce platform, obtaining product information corresponding to each of a plurality of data items of a product registered in the e-commerce platform; obtaining a first index indicating the category of the product based on the product information corresponding to the first data item among the plurality of data items; Obtaining a second indicator indicating the category of the product based on product information corresponding to a second data item among the plurality of data items; Comparing the first indicator and the second indicator, and when the first indicator and the second indicator do not match, specifying that the product information is unclear; An information processing method including the above.

13. An information processing method executed by an information processing device used in an e-commerce platform, Obtaining product information corresponding to each of a plurality of data items of a product registered on the e-commerce platform; Obtaining a first indicator indicating the category of the product based on product information corresponding to a first data item among the plurality of data items; Obtaining a second indicator indicating the category of the product based on product information corresponding to a second data item among the plurality of data items; When the similarity between the first indicator and the second indicator is less than a threshold value, specifying that the product information is unclear; An information processing method including the above.

14. An information processing method executed by an information processing device used in an e-commerce platform, Obtaining product information corresponding to each of a plurality of data items of a product registered on the e-commerce platform; Obtaining a first indicator indicating the probability that the product belongs to a certain category based on product information corresponding to a first data item among the plurality of data items; Obtaining a second indicator indicating the probability that the product belongs to the certain category based on product information corresponding to a second data item among the plurality of data items; Based on the first indicator and the second indicator, specifying that the product information is unclear; An information processing method including the above.

15. A program executed by an information processing device used in an e-commerce platform, The program causes the information processing device to Obtain product information corresponding to each of a plurality of data items of a product registered on the e-commerce platform; Obtain a first indicator indicating the category of the product based on product information corresponding to a first data item among the plurality of data items; Obtain a second indicator indicating the category of the product based on product information corresponding to a second data item among the plurality of data items; Comparing the first index and the second index, and when the first index and the second index do not match, identifying that the product information is unclear; A program for causing the execution. [

16. ] A program executed by an information processing apparatus used in an e-commerce platform, The program causes the information processing apparatus to acquire product information corresponding to each of a plurality of data items of a product registered in the e-commerce platform; obtain a first index indicating the category of the product based on the product information corresponding to a first data item among the plurality of data items; obtain a second index indicating the category of the product based on the product information corresponding to a second data item among the plurality of data items; when the similarity between the first index and the second index is less than a threshold value, identifying that the product information is unclear; A program for causing the execution. [

17. ] A program executed by an information processing apparatus used in an e-commerce platform, The program causes the information processing apparatus to acquire product information corresponding to each of a plurality of data items of a product registered in the e-commerce platform; obtain a first index indicating the probability that the product belongs to a certain category based on the product information corresponding to a first data item among the plurality of data items; obtain a second index indicating the probability that the product belongs to the certain category based on the product information corresponding to a second data item among the plurality of data items; identifying that the product information is unclear based on the first index and the second index; A program for causing the execution.

Images