International Business Machines Corporation

Yuan Jin

Jian Jun Wang

Yin Qian

Xiao Rui Shao

Xi Xi Liu

Fan Xiao Xin

Ke Zhang

Yu Huang

Zhe Yan

Rong Zhao

Xiao Feng Ji

Xiao Cao Cao

Li Li Guan

Li Ping Wang

Jing Jing Zhang

Cheng Yi Huang

Wei Jun Zheng

Cheng Peng

Kai Ye

Dong Huang

Qing Lu

Fang Rui Shu

Ziqiumin Wang

Chengyu Peng

Zheng Ping Ac Chu

Su Liu

Xin Peng Liu

Xin Zhao

Qi Rong Wang

Jing Zhang

Xiao Xia Mao

Zheng Ping A C Chu

Zheng Ping Chu

Video analysis by receiving video data associated with a multi-participant meeting, identifying a first participant viewpoint, a related argument and an authority level of a viewpoint-argument, from the video data using a machine learning model, identifying a first topic within the video data, correlating the first participant viewpoint to the first topic, determining a distance between the first participant viewpoint and a second participant viewpoint, the second participant viewpoint correlated to the first topic, and providing a depiction of a relationship between the first participant viewpoint and the second participant viewpoint according to the distance.

Viewpoint analysis of video data

An approach for assisting software developers to fill the data gap and reproducing client application issue using reported reproducing steps is disclosed. The approach involves constructing service request/response sequence (from the developer's side based on the limited data) from customer's reproducing steps and developer's running inspection result and perform UI-service mapping, redirecting service call to data simulator which replay the service request and response, and adjusting service response with multiple data sources and policies. Lastly, the approach can transform the issue to an automatic test script with the service response in a mockup function (e.g., can be used to extend test data and enhance test cases).

Reproduction client application issue based on layered data replay

Embodiments relate to decoding communications with token sky maps. At least one electronic communication including emoticons having a non-original meaning is received. A candidate meaning is determined for the emoticons having the non-original meaning in the at least one electronic communication based at least in part on token neighborhood distribution structures. The candidate meaning for the emoticons having the non-original meaning is caused to be displayed on at least one device.

Decoding communications with token sky maps

A computer-implemented method includes: extracting, by one or more processors, media data from an information technology ticket; extracting, by one or more processors, a plurality of data elements from the media data; and generating, by one or more processors, a ticket knowledge graph based on the plurality of data elements, wherein a node of the ticket knowledge graph represents a data element, and an edge between a first node and a second node in the ticket knowledge graph represents a correlation between a first data element represented by the first node and a second data element represented by the second node.

Ticket knowledge graph enhancement

A computer-implemented method, system and computer program product for providing high quality speech recognition. A first speech-to-text model is selected to perform speech recognition of a customer's spoken words and a second speech-to-text model is selected to perform speech recognition of the agent's spoken words during a call. The combined results of the speech-to-text models used to process the customer's and agent's spoken words are then analyzed to generate a reference speech-to-text result. The customer speech data that was processed by the first speech-to-text model is reprocessed by multiple other speech-to-text models. A similarity analysis is performed on the results of these speech-to-text models with respect to the reference speech-to-text result resulting in similarity scores being assigned to these speech-to-text models. The speech-to-text model with the highest similarity score is then selected as the new speech-to-text model for performing speech recognition of the customer's spoken words during the call.

Providing high quality speech recognition

Aspects of the present disclosure relate to support content generation. An issue description is received from a user. A software feature associated with the issue description is identified using a trained classification model. A base image associated with the software feature is obtained. The base image is modified to add information indicated in the issue description, wherein the modified base image is generated support content. The generated support content is transmitted.

User support content generation

Aspects of the present disclosure relate to automatically generating a user manual using a technique that includes training a first model with a first set of training data. The technique further includes generating, by the first model, a set of operations and a set of windows, where the set of operations and the set of windows are functions of the program. The technique further includes, generating a plurality of tasks, where a first task comprises a first operation being performed on a first window. The technique further includes determining an order of the plurality of tasks and calculating a level score for the first operation of the first window. The technique further includes assembling the user manual having the plurality of tasks in the determined order.

Intelligent generation and organization of user manuals

An approach is provided in which a system receives a first meme sent from a first user that is targeted to a second user. The system transforms the first meme into a second meme based on analyzing the first meme against a first user profile data corresponding to the first user and a second user profile data corresponding to the second user. The transforming includes replacing at least one element of the first meme with a second element in the second meme. The system, in turn, sends the transformed second meme to the second user.

Meme intelligent conversion

One or more frames within a video are segmented and classified according to information type. At least a portion of the information in the one or more frames is determined to be obscured. In response to this determination, information from audio data and the unobscured portion of the one or more frames is extracted. Using this extracted information, a set of replacement data is generated, according to the information type, for the obscured information in the one or more frames. The set of replacement data and the unobscured portion of the one or more frames is merged into one or more composite frames. These composite frames are displayed to a user.

Composite video frame replacement

Techniques are described herein for transmitting subgroup messages and displaying such subgroup messages within a group-based message interface. The subgroup message includes both a message intended for the subgroup and a subgroup command indicating that the message is sent to a subgroup of a group participating in a group-based message interface. The subgroup members are then determined from the subgroup command and the subgroup message is displayed to the subgroup members.

Subgroup messaging within a group-based messaging interface

Approaches presented herein enable displaying a barrage message. Specifically, one or more objects and location information for each object in a frame of a video are identified. A barrage message to be displayed in the frame of the video is obtained. The barrage message is displayed without covering any object in the frame.

Barrage message processing

A method and system of converting a first language of a soundtrack of a person speaking in a video to a second language. A meaning of a word of the first language is translated, and one or more synonym words in the second language stored in a database of the computer system is located. The first and second languages are different languages. Outlines of shapes of mouth openings of the one or more synonym words in the second language are compared with the outlines of the shapes of mouth openings of the word of the first language. A synonym word of the one or more synonym words translated from the second language into the first language having mouth openings with a smallest difference from the mouth openings of the word of the first language is selected.

Dubbing and translation of a video

The present invention provides a method and system for converting a first language of a soundtrack of a person speaking in a video to a second language. The method defines an outline of a shape of a mouth opening of the person speaking a syllable of a word of the first language in the video at a given start time by selecting a predetermined number of points along a border of the mouth opening defined by the lips. A length of the spoken syllable is also measured and one or more adjacent syllables are combined to create a word. The word is translated into synonym words in the second language, the best fit synonym word is selected that most closely matches the mouth shape of the first language word, and a mouth shape adjustment script is applied to fine tune the mouth shape of the best fit synonym word.

System and method for audio dubbing and translation of a video

This disclosure provides a method and system for pre-processing an E-book, wherein the E-book comprises text pages in scanned format and annotation pages in scanned format and the annotation pages contain annotation items for explaining text elements in the text pages. The method comprises: extracting annotations from the annotation items; identifying annotated elements from the text pages, wherein an annotated element is a text element that may be determined to be associated with one of the annotations; defining positions on the text pages for the annotated elements respectively; and storing the positions and the annotations. The positions and the annotations are correlated according to their relationship with the annotated elements.

Growing community of inventors

Shanghai, China

Yuan Jin