International Business Machines Corporation

Gal Ashour

Reiner Kraft

Arnon Amir

Matthew Denesuk

Brian K Blanchard

Olympia Gluck

Dror Porat

Daniel Nechemia Rotman

Michael Edward Factor

Kenneth Wayne Boyd

Shachar Fienblit

Sam Clark Werner

Eli Malul

Amiram Hayardeny

Ifat Nuriel

Dalit Tzafrir

Yoram Simha Adler

Konstantin Kupeev

Zak Mandel

Naftaly Sharir

Noa Privman-Horesh

Neelakantan Sundaresan

Silvia Melitta Mueller

Aviad Zlotnick

Haggai Roitman

Roy Levin

Mattias Marder

Yoav Medan

Ophir Azulai

Ronald J Lisle

Michal Rimon

Oz Dov Hershkovitz

Dan Chazan

Tal Hakim

Karen Holtz

Avishai Moshe Fedida

Yevgeny Burshtein

Noa Privman Horesh

Ronen Ozer

Amon Amir

A processor may receive a query from a user. The processor may identify one or more subjects in the query. The one or more subjects may include a particular subject. The processor may identify one or more sections of a recording. The processor may group the one or more sections into one or more subdivisions. The processor may determine that the particular subject is in at least one of the one or more sections. The processor may display the one or more subdivisions of the one or more sections that include the particular subject to the user.

Query-based granularity selection for partitioning recordings

Validating microprocessor instruction execution by receiving a floating-point exception selection, receiving a validation method selection, generating validation data according to the floating-point exception selection and the validation method selection by randomly generating a first tensor element value and randomly generating a second tensor element value according to the first tensor element value and the floating-point exception selection, and executing a floating-point computation according to the validation data.

Validating microprocessor performance

A computer implemented method of mapping multimedia analytics of multimedia objects into a resilient distributed dataset (RDD), comprising one or more processors adapted to obtain an RDD of a cluster computing framework executed by a cluster comprising a plurality of computing nodes, the RDD comprises a plurality of entries each comprising a pointer to one of a plurality of multimedia objects stored in a shared storage, instruct each of a plurality of framework tasks executed by at least some members of the cluster to apply a docker operator for retrieving and executing one of a plurality of multimedia containers each associated with a respective one of the multimedia objects and comprising a multimedia processing algorithm for processing the respective multimedia object, receive from the framework tasks multimedia analytics results generated simultaneously by the multimedia containers and map the multimedia analytics results into the RDD.

Multimedia analytics in spark using docker

A system comprising a non-transient computer-readable storage medium having stored thereon instructions and at least one hardware processor configured to execute the instructions, to receive a video sequence; divide the video sequence into one or more scenes based on scene boundaries, wherein each scene comprises a plurality of temporally-contiguous image frames, and wherein said scene boundaries are being determined based on a similarity metric between two temporally-contiguous image frames meeting a dissimilarity threshold; and, for each scene of the one or more scenes, (i) generate a plurality of preliminary classifications of an object appearing in at least some of said image frames in the scene, wherein each of said plurality of preliminary classifications has a confidence score, and (ii) calculate a combined classification of the object based on said plurality of preliminary classifications, wherein each of said preliminary classifications is weighted in accordance with its confidence score.

Object recognition in video

Segmenting serial data by processing multiple candidate segmentation point sets associated with a serial data set, each candidate set including a different number of candidate segmentation points representing the serial data set as segments in accordance with segmentation criteria, where the processing includes determining, for each of the candidate points, a count of the candidate sets that include the candidate point, and creating, for each of the candidate sets, a corresponding alternative segmentation point set associated with the serial data set, where the alternative set includes n alternative segmentation points representing the serial data set as n+1 segments, where n equals the number of candidate points in the candidate set to which the alternative set corresponds, and where the n alternative points in the alternative set correspond to n of the candidate points having the greatest counts, and arranging the alternative sets in order of their numbers of segmentation points.

Non-greedy hierarchical segmentation of serial data

A method comprising using at least one hardware processor for receiving a request of a user for media, retrieving a first metadata set relating to said media, comparing between: (a) a media usage profile of the user, which profile reflects knowledge of the user, and (b) the first metadata set relating to said media, to identify gaps in the knowledge of the user with respect to said media, and providing to the user information relating to said identified gaps.

Predicting knowledge gaps of media consumers

Methods and systems for social analysis in multi-participant meetings provide for receiving a data stream of a multi-participant meeting, wherein multiple topics are discussed. The data stream is analyzed to detect topics and to define topical sub-activities that are performed by at least a portion of the participants. Contributions to a topic by two of the participants in the topical sub-activities are identified and evaluated. A connection weight between the two participants is calculated based on the evaluation of their respective contributions.

Social analysis in multi-participant meetings

A method, system and computer program product are provided for social analysis in multi-party meetings. The method includes receiving a data stream of a multi-participant meeting and identifying participants' contributions in the data stream. The method also includes detecting a sub-activity within the data stream in which two or more participants are active and calculating a connection weight between two participants based on contributions in a sub-activity. Calculating a connection weight between two participants includes increasing a weighting of two participants that interact in the same sub-activity. The method also includes integrating the connection weight between two participants into a weighted social network.

Generating a consistent point in time copy of data in a source volume and a target volume is achieved responsively to a first data modification request by writing a first altered version of the data onto a single source volume, asynchronously transferring the first altered version from the first storage site to a target volume located at a remote second storage site, while avoiding copying the first altered version onto other volumes at the first storage site. While asynchronously transferring the first altered version de-queuing a second modification request, and responsively to the second modification request synchronously transferring the first altered version from the first storage site to the target volume. Then a second altered version of the data is written to the single source volume and a copy transferred to the target volume in like manner.

System, method and computer program product for generating a consistent point in time copy of data

A computing system and method receive a query; separate a plurality of information sources into individual elements of content (EOC); tag each EOC with metadata that indicate source, date, and other relevant information; pattern match each EOC; calculate the respective distance function from every EOC to every other EOC; and output EOC to a set of virtual buffers () containing appropriately related EOC less than a given distance value. The method further creates virtual summary buffers (); then concatenates the EOC in each virtual buffer (); applies a comparative analysis filter () to remove redundant sub-elements; and presents the results as summary digests ().

Synthesizing information-bearing content from multiple channels

A method for generating a consistent point in time copy of data, the method includes: selecting at least one selected data block to be copied from a source volume to a target volume in response to a request to generate a consistent point in time copy of multiple data blocks; waiting until the source volume is ready to send the at least one selected data block to a remote volume while queuing at least one data block modify request; de-queuing the at least one queued modify requests while copying the at least one selected data block from the source volume to the target volume; wherein the copying includes utilizing a first copying mechanism to copy a first selected data block if a request to modify the first selected block is de-queued before the first selected data block is copied to the target volume; else, the copying includes using a second copying mechanism that is slower than the first copying mechanism.

A method, system and computer program product for copying data are provided. The method includes: establishing a virtual copying relationship at a current primary site; and establishing a real copying relationship with a first remote site in response to a reception of a first remote site information by the current primary site information while maintaining the virtual copying relationship.

System, method and computer program product for copying data

Convergent playback of buffered content to real-time content play. A content viewing system is provided including a content player operative to receive a content stream from a content source, a content viewer operative to receive the content from the content player, and a buffer into which the content player may store the received content and from which the player may provide the stored content to the content viewer, where the content player is operative to buffer the content stream to the buffer responsive to a request to pause play of the content stream, and retrieve the content from the buffer for playback of the buffered content stream responsive to a request to resume play of the content stream, thereby emptying the buffer, where the buffer is emptied at a faster rate than the rate at which the content stream is buffered, and the rate at which the buffer is emptied varies.

Convergent playback of buffered content to real-time content play

A method receives a query; separates a plurality of information sources into individual elements of content (EOC); tags each EOC with metadata that indicate source, date, and other relevant information; pattern matches each EOC; calculates the respective distance function from every EOC to every other EOC; and outputs EOC to a set of virtual buffers () containing appropriately related EOC less than a given distance value. The method further creates virtual summary buffers (); concatenates the EOC in each virtual buffer (); applies a comparative analysis filter () to remove redundant sub-elements extracts context-preserving EOC, determined based upon a respective degree of corroboration of the context preserving EOC, from each respective concatenated EOC. The method synthesizes the content-preserving EOC extracted from each respective concatenated EOC into summary digests; and presents the results as summary digests ().

Growing community of inventors

Yokneam, Israel

Gal Ashour