International Business Machines Corporation

Hcl Technologies Limited

Jonathan Frederick Brunn

Jeffrey Robert Hoy

Asima Silva

Marit L Imsdahl

Jennifer E Heins

Jessica Whitley Forrester

Stephen Carlyle Hess

Kulvir Singh Bhogal

Jonathan D Dunne

Ethan A Geyer

Marc D Labrecque

Ami H Dewar

Amy Widmer

Rui Zhang

Bo Jiang

Rachael M H Dickens

Scott E Chapman

Vincent P Burckhardt

Ryan Alexander Davis

Rogelio Vazquez-Rivera

Naama Tepper

Dennis Jinlone Chen

Rachael Marie Huston Dickens

Claudio Procida

Robert W Campion

Liam S Harpur

Heiko Hary Ludwig

Christopher D Johnson

Andrew T Penrose

Razeyah Stephen

John Girata, Jr

Rachael M Dickens

Jonathan T Coffey

Mark A DuFresne

James J Stadtmiller

Yuriy B Veytsman

Laura Janet Rodriguez

Marshall Allen Lamb

Yi Zhou

Patrick Michael Commarford

Erika Varga

Stephen John Foley

Kelley M Gordon

Michael Celedonia

Troy Anthony Howard

Katelyn Applegate

Yu Zhuang

Sheng Hu

Yuan Cheng

Ming Wan

Thomas J Blanchflower

Bo Min Nie

Melissa N Fritcher

Matthew S Hunt

Kelvin T Ly

David P Zmick

Clayton P Coleman

Sarah B Lips

Gayle C Thiel

Embodiments relate to an intelligent computer platform to build a federated learning framework including creating a hierarchy of machine learning models (MLMs). The hierarchy of MLMs has a primary MLM in a primary layer. Training the primary MLM includes capturing contributing model updates across at least one communication channel. A secondary MLM is created and logically positioned in a secondary layer of the hierarchy. The secondary MLM is operatively coupled to the primary MLM across the at least one communication channel. The created secondary MLM is initialized, including cloning weights and framework of the primary MLM into the secondary MLM, and populated with secondary data. The populated data has model updates local to the created secondary MLM. The secondary MLM is logically stored local to the secondary layer, and limits access to the secondary MLM to the secondary layer.

Building a federated learning framework

Provided are a computer program product, system, and method for generation of tasks and retraining machine learning modules to generate tasks based on feedback for the generated tasks. A machine learning module processes an input text message sent in the communication channel to output task information including an intended action and a set of associated users. A task message is generated including the output task information of a task to perform. The task message is sent to a user interface panel in a user computer. Feedback is received from the user computer on the output task information in the task message. The machine learning module is retrained to output task information from the input text message based on the feedback to reinforce likelihood correct task information is outputted and reinforce lower likelihood incorrect task information is outputted.

Automatic generation of tasks and retraining machine learning modules to generate tasks based on feedback for the generated tasks

Using a natural language analysis, a current message is classified into a current message class, the current message being a portion of an interaction in narrative text form. Using a trained message class prediction model, a probability of a previous message class having resulted in the current message class is determined. A previous message is extracted from the interaction using the probability, the previous message being a portion of the interaction occurring prior to the current message, the previous message being classified into the previous message class.

Identifying related messages in a natural language interaction

A method, system and computer program product for improving understanding of comments on collections of data. A social media stream is monitored for comments to a container (e.g., photo album) and elements within the container (e.g., photographs). These comments are stored in a data structure along with an identification of the container or element of the container upon which the comments are directed. In response to a user selecting to view comments to an element within the container or the container itself, the data structure is searched for the comments to the selected element or container. Images of the container are then displayed in a fliptych manner, where the selected element or container is displayed in the center section of the fliptych. A list of comments is displayed in a comments section below the fliptych, where the list of comments includes highlighted comments directed to the selected element or container.

Understanding of the relationship between the comments being made to the containers and the comments being made to the elements of the containers

An in-context cognitive information assistant is provided by: obtaining a context for a user, wherein the context comprises a calendar activity with one or more other users; supplementing the context by obtaining one or more conversations with the user related to the context; extracting cognitive data for the context and the conversations; and finding relevant materials in a corpus using the cognitive data. The relevant materials are used to prepare the user for interactions with other users.

In-context cognitive information assistant

Using a natural language analysis, a current message is classified into a current message class, the current message being a portion of an interaction in narrative text form. For the interaction using a state prediction model, an interaction outcome corresponding to the current message class is forecasted, the forecasting comprising computing a probability that the current message class will result in a successful message class. Using the state prediction model, a set of next message classes and a set of predicted interaction outcomes are determined, each message in the set of next message classes corresponding to the current message class, each predicted interaction outcome in the set of predicted interaction outcomes corresponding to a next message class in the set of next message classes. According to the corresponding predicted interaction outcome, the set of next message classes is ranked.

Outcome of a natural language interaction

Identification of content classes deemed important by users is received. Classifiers are trained on user communication text to recognize content matching content classes. A composite classifier is applied to text of communication among users of a second set of users, including tagging, by the composite classifier, content in text of communication among the users of the second set. The tagging is responsive to the classifiers recognizing their respective content classes in text of communication among the users of the second set.

Composable natural language lenses for collaborative streams

Controlling access to categorized messages includes categorizing a message into a number of categories according to a message objective. The message objective informing user association, message association, and message access rules for each of the number of categories. Controlling access to categorized messages includes determining, based on the message objective, a number of users allowed access to the message. Controlling access to categorized messages includes allowing the number of users to access the message according to the message objective.

Determining a number of users allowed to access message based on message objective

A method, a device and a computer program product for processing a segment are proposed. In the method, a property of at least one of a first segment and a second segment in a segment set is obtained. The segment set includes a plurality of segments belonging to at least one conversation. The second segment occurs after the first segment. A boundary feature of at least one of the first segment and the second segment is determined based on the property. The boundary feature indicates whether there is a boundary of a conversation after the first segment.

Conversation boundary determination

Familiarity-based text classification framework selection is described. A list of participants in an electronic message thread is selected. For each pairing of participants, a familiarity score is determined based on a number of criteria. A familiarity model is formed based on multiple familiarity scores and a text classification framework for the electronic message thread is selected based on the familiarity model.

Familiarity-based text classification framework selection

Provided are techniques for finding content on computer networks. A search request having search terms is received. The search request is executed against an index to identify first search results, wherein the index comprises mappings of search terms to web sites and mappings of search terms to patterns. In response to determining that the search request failed based on first search results, an unknown search term in the search terms that does not exist in the index is found. The unknown search term is matched to a pattern of a similar search term using the mappings of search terms to patterns. A candidate web site is identified based on the matching. Additional content is requested from the candidate web site for the unknown search term to identify second search results comprising previously hidden content. The additional content is indexed.

Finding content on computer networks

An indicator associated with a financial card to tender payment for a transaction is received from a point of sale device. A storage device stores independent sets of information associated with the financial card. The sets of information indicate different actions to be performed and settings corresponding to the actions. The indicator received from the point of sale device is associated with one of the independent sets of information. The actions associated with the received indicator are performed in accordance with the corresponding settings in response to conducting a transaction with the financial card.

Conducting various actions indicated by a financial card

A method, system and computer program product for handling notifications. A notification system monitors content (e.g., photograph) followed by a user for updates on a communication system. If the notification system identifies updates to content out of the monitored content, then the notification system applies the policy for delivering notifications concerning the content followed by the user that was updated. The notification system registers to receive information about updates to the content followed by the user as well as new events on the communication system. In response to identifying updates to content related to the content followed by the user that was updated from the received information, a summary concerning the updates to the content followed by the user and the related content is then generated. A notification is then delivered to the user containing the generated summary in a manner consistent with the policy.

Intelligently delivering notifications including summary of followed content and related content

Growing community of inventors

Logan, UT, United States of America

Jonathan Frederick Brunn