International Business Machines Corporation

Kyndryl, Inc.

Gray Franklin Cannon

Aaron K Baughman

Shikhar Kwatra

Stephen Carl Hammer

Marco Aurelio Stelmar Netto

Vagner Figueredo De Santana

Gary William Reiss

Craig M Trim

Nicholas Michael Wilkin

Raghuveer Prasad Nagar

Christian Eggenberger

Indervir Singh Banipal

Christopher Eythan Holladay

Micah Forster

Corey B Shelton

Sara Perelman

Sarbajit Kumar Rakshit

Vijay Ekambaram

Mauro Marzorati

Vijay Kumar Ananthapur Bache

Saravanan Sadacharam

Mary E Rudden

Jennifer L Szkatulski

Ryan Lynch Whitman

David Alexander Provan

Ashraf Ali Thajudeen

Mohit Trivedi

Patrick Veedock

Sheldon Busby

Eris Opal Rashon Calhoun

By analyzing a natural language content concurrently with a presentation of the natural language content, a first language of the natural language content is detected. By analyzing the natural language content concurrently with the presentation of the natural language content, it is detected that a subset of the natural language content is expressed in a second language, wherein the second language is different from the first language, wherein a set of known languages comprises the first language, wherein the set of known languages excludes the second language. From the subset of the natural language content, a modified subset expressed in the first language is generated. The modified subset is inserted into the natural language content.

Contextual and personalized real time content modification for language assistance

An embodiment includes identifying, from among the plurality of digital content datasets, a set of candidate textual items based on relevance to a specified subtopic using one or more natural language processing techniques. The embodiment groups candidate textual items into a predetermined number of groups using relevance scores and feature vectors. The embodiment trains a pre-trained encoder-decoder model using a designated group of selected textual items, where the pre-trained encoder-decoder model is pretrained to generate textual content according to a particular style of writing. The embodiment generates, using the pre-trained encoder-decoder model, machine-authored textual content in the particular style of writing resulting in an article about the specified subtopic based on the designated group of selected textual items. The embodiment also includes transmitting the article to a remote web server as an update for a website.

Elucidated natural language artifact recombination with contextual awareness

Using a classical data model executing on a classical processor, a set of classical features is scored. A score of a classical feature comprises an evaluation of a utility of the classical feature in predicting a result involving a resource. Using a quantum data model executing on a quantum processor and the scored set of classical features, a set of quantum features is scored. The quantum data model is executed a number of times previously determined using a set of results of executing the quantum data model on a set of annotated training data. The scored set of classical features and the scored set of quantum features are correlated, forming a combined set of scored features. Using the combined set of scored features and a first set of input data of a resource, a valuation of the resource is calculated.

Identifying related messages in a natural language interaction in multiple iterations

A customized video stream is provided. A second content is searched for that includes the plurality of entities, and a plurality of user data associated with the social media account. A hierarchy of additional content is generated based, at least in part, on the second content that is associated with the plurality of entities, and the user data associated with the social media account, wherein the hierarchy of additional content contains a plurality of second content. A customized video stream is generated based, at least in part, on the video stream and the hierarchy of additional content. The customized video stream is populated on a display screen of the user device, wherein the video stream is displayed on the display screen of the user device, and the hierarchy of additional content is displayed in a portion of the display screen and displayed concurrently with the video stream.

Video annotation based on social media trends

A method, system, and computer program product for implementing machine learning natural language digital package generation and selection is provided. The method includes receiving from hardware and software sources, factual data associated with an event. In response, natural language digital templates comprising natural language phrase variants for each portion of the factual data is generated. Factual data phrases are generated and packaged into digital packages including at least one natural language phrase variant with respect to each portion of factual data. An initial package is selected by minimizing a number of repetitions of the factual data phrases across the digital packages and digital summaries are extracted. Alignment attributes associated with the digital summaries are determined with respect to the initial package and a final package is selected. A hardware device is enabled for presenting a video stream including the final package with respect to the event.

Automated natural language generation and selection

Modifying digital content based on predicted future user behavior is provided. Trends in propagation values corresponding to a layer of nodes in an artificial neural network are identified based on measuring the propagation values at each run of the artificial neural network. The trends in the propagation values are forecasted to generate predicted propagation values at a specified future point in time. The predicted propagation values are applied to the layer of nodes in the artificial neural network. Predicted website analytics values corresponding to a set of website variables of interest for the specified future point in time are generated based on running the artificial neural network with the predicted propagation values. A website corresponding to the set of website variables of interest is modified based on the predicted website analytics values corresponding to the set of website variables of interest for the specified future point in time.

Deep forecasted human behavior from digital content

Synchronizing a sensor network and an ontology by analyzing outputs of a set of sensors, analyzing relationships of an ontology, mapping the outputs to the relationships, identifying a correlation among the outputs, and modifying the ontology according to the correlation.

Synchronizing a sensor network and an ontology

A method, computer system, and a computer program product for speech synthesis is provided. The present invention may include generating one or more final voiceprints. The present invention may include generating one or more voice clones based on the one or more final voiceprints. The present invention may include classifying the one or more voice clones into a grouping using a language model, wherein the language model is trained using manually classified uncloned voice samples. The present invention may include identifying a cluster within the grouping, wherein the cluster is identified by determining a difference between corresponding vectors of the one or more voice clones below a similarity threshold. The present invention may include generating a new archetypal voice by blending the one or more voice clones of the cluster where the difference between the corresponding vectors is below the similarity threshold.

Voice cloning transfer for speech synthesis

From a set of natural language text documents, a concept tree is constructed. For a node in the concept tree a polarity of the subset represented by the node is scored. A second set of natural language text documents is added to the subset, the adding resulting in a modified subset of natural language text documents having a polarity score within a predefined neutral polarity score range. From the modified subset, a bin of sentences is selected according to a sentence selection parameter, a sentence in the bin of sentences being extracted from a selected document in the modified subset. A sentence having a factuality score below a threshold factuality score is removed from the bin of sentences. From the filtered bin of sentences a new natural language text document corresponding to the filtered bin of sentences is generated using a transformer deep learning narration generation model.

Automatic neutral point of view content generation

An embodiment associates a user's post data that with a category from a predefined list of categories based on the post content. The embodiment analyzes, using machine learning, the post data for potentially sensitive content and generates a first sensitive data indicator identifying potentially sensitive information in the post data and an associated first confidence value. The embodiment generates explanatory data identifying a feature that contributed to the post data being identified as potentially sensitive, and generates a modified version of the post data that modifies the feature. The embodiment analyzes the modified post data for potentially sensitive content and generates a second sensitive data indicator and a second confidence value indicating that the post data is more likely to contain sensitive data than the modified post data. The embodiment alerts the user regarding the potentially sensitive data and recommends changing the post based on the modified feature value.

Recommending post modifications to reduce sensitive data exposure

A computer-implemented method, system and computer program product for generating test cases for testing software. Log files are analyzed for indications of software defects resulting in software errors. For those log files that were found to indicate software errors, such log files are analyzed to determine if they contain personal data that needs to be anonymized based on an entropy score of the personal data. When the entropy score of the personal data is below a threshold value, anonymized variations of the personal data are generated. Trace logs are then created containing the anonymized variations of the personal data, which are returned to a software tester to create a test case for testing software using the created trace logs without having access to personal data. In this manner, the software tester may reproduce the software errors without having access to the personal data.

Creating test cases for testing software using anonymized log data

Provided is a method, a system, and a program product for determining a policy using semi-supervised reinforcement learning. The method includes observing a state of an environment by a learning agent. The method also includes taking an action by the learning agent. The method further includes observing a new state of the environment and calculating a reward for the action taken by the learning agent. The method also includes determining whether a policy related to the learning agent should be changed. The determination is conducted by a teaching agent that inputs the state of the environment and the reward as features. The method can also include changing the policy related to the learning agent upon a determination that a label outputted by the teaching agent exceeds a reward threshold.

Semi-supervised reinforcement learning

According to a first aspect of the present invention, there is provided a computer implemented method, a computer system and a computer program product, including training a set of exploitation models, training a set of exploration models, generating a combined exploitation and exploration heat map, and inputting the combined exploitation and exploration heat map into a convoluted neural network.

Deep evolved strategies with reinforcement

A system for executing a video summary is provided. One or more video segments for a video based on one or more digital media is generated. A video summary is generated based on a user request.

Growing community of inventors

Atlanta, GA, United States of America

Gray Franklin Cannon