International Business Machines Corporation

Daisuke Takuma

Emiko Takeuchi

Hirobumi Toyoshima

Yoshinori Kabeya

Issei Yoshida

Hiroki Yanagisawa

Yuta Tsuboi

Takuma Murakami

Yuki Makino

Hiroki Oya

Kohichi Kamijoh

Shoko Suzuki

Masaki Ono

Tetsuya Nasukawa

Yohei Ikawa

Hiroshi Kanayama

Shunichi Amano

Hiroshi Nomiyama

Shinsuke Mori

Ryosuke Kohita

Kinya Kuriyama

Katsumasa Yoshikawa

Mariko Nagai

Noriyuki Ikawa

Frequent sequences extracted from a set of documents according to a common rule are obtained. Based on comparing occurrence frequencies of various sequences, confidence of the first frequent sequence being a label expression representing a document part in a target document is evaluated. Keywords are extracted from the target document based on evaluation of the confidence.

Text mining based on document structure information extraction

A computer-implemented method is provided for suggesting a query term. The method includes obtaining a seed term and a candidate term. The method further includes retrieving two nodes indicating the seed term and the candidate term respectively from a semi-structured corpora. The method also includes determining, by a processor device, whether the candidate term includes details of information on the seed term based on a positional relation between the two nodes in the semi-structured corpora. The method additionally includes suggesting the candidate term as an additional query term in response to a positive determination. The method further includes performing the query with the candidate term as the additional query term in response to a suggestion of use acceptance to enhance query answer retrieval results.

Suggesting query terms

The present invention may be a method, a computer system, and a computer program product for suggesting an output candidate. The method comprises receiving a user input; selecting a corpus containing an expression similar to the user input among a plurality of corpuses; finding, in the user input, a seed word that may be present in a definition statement of an entry in a dictionary; identifying, in the dictionary, an entry of a definition statement containing the seed word or within a threshold similarity to the seed word with reference to the selected corpus; and suggesting the identified entry as an output candidate.

Suggestion of an output candidate

A fast neural transition-based parser. The fast neural transition-based parser includes a decision tree-based classifier and a state vector control loss function. The decision tree-based classifier is dynamically used to replace a multilayer perceptron in the fast neural transition-based parser, and the decision tree-based classifier increases speed of neural transition-based parsing. The state vector control loss function trains the fast neural transition-based parser, the state vector control loss function builds a vector space favorable for building a decision tree that is used for the decision tree-based classifier in the neural transition-based parser, and the state vector control loss function maintains accuracy of neural transition-based parsing while the decision tree-based classifier is used to increase the speed of the neural transition-based parsing while using the decision tree-based classifier to increase the speed of the neural transition-based parsing.

Computation reduction using a decision tree classifier for faster neural transition-based parsing

A search query is received from a search engine of multiple search engines. One or more search results from each search engine of the multiple search engines in response to the search query are received. Each search result of the one or more search results includes a first search score. A factor is determined for each search result of the one or more search results. A second search score is determined for each search result of the one or more search results based on the first search score and the factor. The one or more search results are displayed to a user of the search engine in an order. The order is based on the second search score.

Collaborative search of databases

A method, a system, and a computer program product for predicting an outcome expected for a particular positional value is provided. In the method, an input set of data records, each having a label and a positional value, and a target positional value are obtained. The label of each data record is one in a label set. A learning model that includes an output layer, an input layer corresponding to the label set and a network structure provided therebetween is read. In the learning model, the network structure has a plurality of functions trained so as to evaluate influence from each label in the label set depending on a relationship between the target positional value and a representative positional value associated with the label in the label set. A target outcome is estimated for the target positional value from the input set using the learning model.

Regression for metric dataset

A computer retrieves a dialog information records of the active call of the first operator. The computer extracts features from the dialog information records. The computer determines a feature vector from the extracted features and determines a transfer probability value based on the feature vector and previous call transfers to the second operator.

Call transfer support system

A method of creating a presentation includes extracting text parts from content, assigning first attributes to the text parts based on the content and design similarities between the text parts, assigning second attributes to the text parts based on processing the text content to extract contextual similarities between the text parts, determining groups of the text parts based on the first and second attributes, determining first links between the text parts of different groups using the display information and adjacencies of the text parts, determining second links of the text parts to object types defined by a graph restriction listing the object types and their parameters, and outputting a presentation by setting the text parts in the presentation corresponding to the parameters of their linked object type in the graph restriction, wherein the text parts of each of the groups are associated with a same one of the object types.

Automated creation and design of presentation charts

Embodiments relate to a type of expression based on a particular theme. An aspect includes acquiring, by an electronic apparatus, from text data for learning, a subset of the text data associated with the particular theme and with particular time period information. Another aspect includes extracting text data containing negative information from the acquired subset of the text data. Another aspect includes extracting a word or phrase having a high correlation with the extracted text data or a word or phrase having a high appearance frequency in the extracted text data from the extracted text data. Yet another aspect includes determining that the extracted word or phrase is the type of expression based on the particular theme.

Detecting expressions learned based on a theme and on word correlation and co-occurence

Methods and systems for searching for similar documents include comparing an input index of a requested document to one or more stored indices for respective stored documents to produce a similarity score for each of the stored documents. Each index indicates which of a plurality of queries matched a respective document. The stored documents are filtered to remove dissimilar documents based on a comparison of each respective similarity score to a threshold. A list of any stored documents that remain after said filtering is output.

Fast filtering for similarity searches on indexed data

A first element is extracted from a pair including a past inquiry and a past response, wherein the first element indicates that the past response shows an understanding of the past inquiry. A model is generated used to estimate a second element in a new inquiry based on the first element, wherein the second element indicates that a new response to the new inquiry shows an understanding of the new inquiry.

Supporting generation of a response to an inquiry

A method and system are provided for collecting social media users who have a specific profile. The method includes retrieving a set of lists connected by at least one criterion to a particular list that is included in a set of reliable lists whose users have already been reliably deemed to have a specific profile. The method includes calculating a list name based confidence value and a list member based confidence value for each list in the retrieved set of lists. The method includes updating the set of reliable lists by adding all lists in the retrieved set of lists that have the list name based confidence value above a first threshold value and the list member based confidence value above a second threshold value. The method includes outputting a listing of users belonging to set of reliable lists as the social media users who have the specific profile.

Collecting social media users in a specific customer segment

A keyword to be categorized is received. A category dictionary including categories having associated registered keywords, and a text corpus are received. Registered keywords are identified in the category dictionary having a degree of similarity to the keyword to be categorized that is equal to or greater than a predetermined value, and the categories associated with the identified registered keywords are extracted. Registered keywords are identified that are co-occurring in the text corpus with the keyword to be categorized, and the categories associated with the identified co-occurring registered keywords are extracted. A degree of importance is determined for each extracted category based on a function of the identified registered keywords in the category dictionary and/or a function of the identified co-occurring registered keywords. The extracted categories are outputted, with at least an indication of each category's relative importance, as category candidates for categorizing the keyword to be categorized.

Growing community of inventors

Tokyo, Japan

Daisuke Takuma