International Business Machines Corporation

Paul Christesten Castro

Marco Pistoia

John J Ponzo

Archan Misra

Umut Topkara

Norman Howard Cohen

Frederique Anne Giraud

Yun-Wu Huang

Apratim Purakayastha

Danny L Yeh

Yunwu Huang

Ravindranath Konuru

Gregory Louis Truty

Gegi Thomas

Peter Kissa

Mauro Baluda

Bishwaranjan Bhattacharjee

Chatschik Bisdikian

David M Lubensky

Joseph W Ligman

James Thomas Rayfield

Vinod Muthusamy

Aleksander A Slominski

Avraham Leff

Marion Lee Blount

Maria Rene Ebling

Vatche Isahagian

John Sidney Davis, Ii

Wolfgang Segmuller

Barry Leiba

Ravi B Konuru

Ji Young Roe

Jinwon Lee

Caroline R Canning

Gina M Cardosi

Lucy L Chen

Patrick M Clough

Matt Harper

John J Mutter

Danny Lo Tien Yeh

Techniques for distributing the training of machine learning models across a plurality of computing devices are presented. An example method includes receiving, from a computing device in a distributed computing environment, a request for a set of outstanding jobs for training part of a machine learning model. A system transmits, to the computing device, information identifying the set of outstanding jobs. The system receives, from the computing device, a selected job for execution on the computing device from the set of outstanding jobs. A chunk of training data associated with the selected job and one or more parameters associated with the selected job may be transmitted to the computing device, and the system may take one or more actions with respect to the chunk of data associated with the selected job based on a response from the computing device.

Peer assisted distributed architecture for training machine learning models

A group of remote devices executing an omnichannel application are coordinated from a network node. An omnichannel mediator coordinates formation of at least two of said remote devices into an omnichannel cloudlet. A component manager controls which of a plurality of components of said omnichannel application should optimally be placed on which individual devices of said omnichannel cloudlet and how data should flow to individual devices of said omnichannel cloudlet. A replication optimizer optimally coordinates data replication for the group of remote device.

Omnichannel approach to application sharing across different devices

A hospital discharge performance system and computer-implemented method for integrating and recording both planned clinical activity and non-clinical barriers to patient discharge. One or more non-clinical barriers are received from at least one remote input device configured to record barriers in close proximity to the occurrence of the barrier and stored in the a non-clinical barrier database. Clinical tasks that need to be performed prior to discharge are integrated with non-clinical barriers that must be overcome to provide an indication of at least one patient's current status with respect to discharge. Update information indicating progress toward completion of planned clinical activity and progress toward overcoming the received one or more non-clinical barriers is received from at least one remote input device. The occurrences of barriers are summarized across a selected group of patient discharges and a report is generated.

Task management tool for patient discharge

Systems for proximity-based access control include a proximity module configured to determine whether a distance from a first mobile device to each of one or more safe mobile devices falls below a threshold distance; a policy engine comprising a processor configured to determine whether a number of safe mobile devices within the threshold distance exceeds a safe gathering threshold; and a security module configured to activate a safe gathering policy in accordance with the safe gathering threshold that decreases a security level in the first mobile device.

Social and proximity based access control for mobile applications

Methods and systems for proximity-based access control include determining whether a distance from a first mobile device to each of one or more safe mobile devices falls below a threshold distance; determining whether a number of safe mobile devices within the threshold distance exceeds a safe gathering threshold with a processor; and activating a safe gathering policy in accordance with the safe gathering threshold that decreases a security level in the first mobile device.

A method for tracking provenance information includes catching and logging data events performed by an instrumented application at a provenance layer. Overwritten library call instructions provide library calls to one or more instrumented libraries that invoke the provenance layer to track data operations. A provenance log is created that includes the logged data events.

Transparently tracking provenance information in distributed data systems

Systems for enabling an application to track provenance include an instrumented application binary on a client device. Overwritten library call instructions provide library calls to one or more instrumented libraries that invoke a provenance layer to track data operations. The provenance layer is configured to catch and log data events performed by the instrumented application.

Systems and methods for providing content using dynamic action templates are provided. A method for providing content to at least one user, comprises corresponding at least one action to at least one action template, associating the content with the at least one action, and presenting the content to the at least one user, wherein the at least one action template is appended to the content.

Providing content using dynamic action templates

A method for sharing documents includes initiating a sharing session with a first participant and a second participant, wherein the first participant is communicatively connected to the second participant via a server, receiving a control request having an action from the first participant, determining whether the control request includes a share page action, and retrieving components of the page with available component types and sending components associated with preferences of the first participant to the first participant and components associated with preferences of the second participant to the second participant responsive to determining that the control request includes the share page action.

Sharing of documents with semantic adaptation across mobile devices

Growing community of inventors

Sharon, MA, United States of America

Paul Christesten Castro