International Business Machines Corporation

Globalfoundries Inc.

Blaine H Dolph

Sandeep R Patil

Riyazahamad Moulasab Shiraguppi

Gandhi Sivakumar

Matthew B Trevathan

David M Lubensky

Nataraj Nagaratnam

Justin David Weisz

Mal Pattiarachi

Miku K Jha

Marco Pistoia

Nitendra Rajput

Ramamohan Chennamsetty

Christopher Joseph Gansen

Aaron K Baughman

Dean Hildebrand

John R Dingler

Raman Harishankar

Palani Sakthi

Ram Viswanathan

Susan L Sylvia

Jermaine Charles Edwards

Sumier Phalake

Kelly E Songer

Frederick H Balboni

Catalin Sava

Kamran R Khan

Carlos A Paez, Jr

Dario R Abramskiehn

Michael Bender

Robert J Torres

John Falk Kelley

Ching-Yung Lin

Todd P Seager

Jui-Hsin Lai

Jennifer Martin

Marcus D Roy

Gennaro A Cuomo

Ira L Allen

John A Blanchard, Iii

John Nathan Tolva

Shamayun Miah

Christopher Hay

Hitesh V Khanna

Stephanie Black Anderson

Computer-implemented methods for a large language model privacy preservation system. Aspects include receiving prompt data from a user device. Aspects further include generating pre-processed prompt data using the prompt data from the user device. Aspects also include identifying a category for the pre-processed prompt data using topic modeling. Aspects include generating normalized prompt data using the pre-processed prompt data. Aspects further include storing the category and the normalized prompt data.

Large language model privacy preservation system

Disclosed embodiments provide a computer-implemented technique for monitoring deviation from a meeting agenda. A meeting moderator and meeting agenda are obtained. Meeting dialog, along with facial expressions and/or body language of attendees is monitored. Natural language processing, using entity detection, disambiguation, and other language processing techniques, determines a level of deviation in the meeting dialog from the meeting agenda. Computer-implemented image analysis techniques ascertain participant engagement from facial expressions and/or gestures of participants. A deviation alert is presented to the moderator and/or meeting participants when a deviation is detected, allowing the moderator to steer the meeting conversation back to the planned agenda.

Cognitive scribe and meeting moderator assistant

A file division and erasure code application executing in a controlling computational device generates data fragments and parity fragments of a file. Each of the generated data fragments and parity fragments are distributed in a different distributed computational device of a plurality of distributed computational devices, where distributing of a generated data fragment comprises determining whether any distributed computational device already stores a duplicate copy of the generated data fragment, and if any distributed computational device already stores the duplicate copy of the generated data fragment, placing the generated data fragment in the distributed computational device if no other data or parity fragment of the file has already been placed in the distributed computational device.

Placement of data fragments generated by an erasure code in distributed computational devices based on a deduplication factor

A method, computer program product, and a system where a processor(s) of a computing device identifies at least two displays coupled to the processor(s) to render visual objects to a user of the computing device, like a cursor, and processor(s) obtains instructions to render the cursor in various locations on the displays via an input device. The processor(s) obtain a display layout for the displays and resolutions of displays and generates an initial virtual desktop that includes physical real estate of the displays where the processor(s) renders visual objects. The processor(s) determines boundaries of each display and selects outermost two dimensional boundaries of the initial virtual desktops and updates the initial virtual desktop to an area bounded by a virtual rectangle comprising the outermost boundaries, which includes the displays and physical space not useable by the processor(s) to render visual objects.

Providing visual control guidance on a display

Aspects of the present invention disclose a method for providing access to hyperlinks on-demand in real-time from a livestream. The method includes one or more processors determining a set of coordinates on a display of a computing device corresponding to an interaction from a user with a video presentation. The method further includes extracting a subset of pixels from an image of the display of the computing device that corresponds to the set of coordinates. The method further includes extracting a uniform resource locator (URL) from the subset of pixels of the image. The method further includes validating the extracted URL from the subset of pixels of the image. The method further includes providing content corresponding to the extracted URL from the subset of pixels of the image to the user.

Accessing embedded web links in real-time

A mobile app is analyzed to determine each UI configuration in a set of UI configurations that is reachable in a possible use of the mobile app. A set of elements used in the mobile app relative to a selected UI configuration from the set of UI configurations is detected. A function performed by a selected element relative to the selected UI configuration is analyzed. The function is matched with a function identified in a hierarchy of elements, the hierarchy being applicable to a plurality of mobile apps. A category of the function identified in the hierarchy is assigned to the selected element. Using the hierarchy, a type of development skill used in developing the selected element is determined. In a blueprint of the mobile app, the category, the type of the development skill, and a numerosity of occurrences of the selected element in the mobile app are output.

Creating an on-demand skills blueprint of a mobile application

A method, computer program product, and a system where a processor(s) of a computing device identifies at least two displays coupled to the processor(s) to render visual objects to a user of the computing device, like a cursor, and processor(s) obtains instructions to render the cursor in various locations on the displays via an input device. The processor(s) obtain a display layout for the displays and resolutions of displays and generates an initial virtual desktop that includes physical real estate of the displays where the processor(s) renders visual objects. The processor(s) obtains a request and generates a simulacrum of the displays, in virtual space. The processor(s) displays, on a portion of the virtual desktop, the simulacrum. The processor(s) obtains, via the input device, a selection of a location on the simulacrum and automatically moves the cursor to a location on the virtual desktop that corresponds to the selected location.

Input placement control guidance on a display

A method and associated systems for a deduplication module of a database-management system. The database-management system, upon receiving a request to perform a transaction that will revise a stored database record, uses memory-resident logs and previously generated database-maintenance tables to: i) identify a first logical block that identifies an updated value, stored in a physical block of storage, to be used to update the database record; and ii) further identify a second logical block that stores in the database a corresponding existing value of the same record. After determining that the first and second logical blocks reside on physical storage devices within the same storage tier, the system directs the deduplication module to associate both logical blocks with the first physical block.

High-efficiency deduplication module of a database-management system

A hierarchal storage management method is provided. The method includes detecting a first portion of a first file being deleted from a hybrid storage device including a hard disk drive (HDD) memory device, a solid state drive (SSD) memory device, and an archival storage memory device. A first set of memory blocks associated with the first portion of the first file is identified. The first set of memory blocks are determined to reside on the SSD memory device. In response, the first set of memory blocks are transferred from the SSD memory device to a first portion of the hybrid storage device.

Tier based data file management

Techniques are disclosed for generating a voice user interface (VUI) modality within an application that includes graphical user interface (GUI) screens. A GUI screen parser analyzes the GUI screens to determine the various navigational GUI screen paths that are associated with edge objects within multiple GUI screens. Some edge objects are identified as select objects or prompt objects. A natural language processing system generates a select object synonym data structure and a prompt object data structure that may be utilized by a VUI generator to generate VUI data structures that give the application VUI modality.

Automated voice enablement of applications

Techniques for optimizing training data within voice user interface (VUI) of an application under development are disclosed. A VUI feedback module synthesizes human speech of a training phrase. This phrase is presented upon a speaker which is simultaneously captured upon a microphone. A speech to text framework converts the synthesized training phrase into text (textualized training phrase). The VUI feedback module compares the textualized training phrase to the actual training phrase and generates a speech training data structure that identifies similarities or dissimilarities between the textualized training phrase and the actual training phrase. This data structure may be utilized by an application developer computing system to identify training data that is most venerable to misinterpretation when a user interacts with the VUI. The VUI may subsequently be adjusted to account for the vulnerabilities to improve operations or user experience of the VUI.

Training data optimization in a service computing system for voice enablement of applications

Training data optimization for voice enablement of applications

An analysis of a mobile app is performed to determine each UI configuration in a set of UI configurations that is reachable in a possible use of the mobile app. A set of elements is detected that is used in the mobile app relative to a selected UI configuration from the set. A function performed by a selected element relative to the selected UI configuration is analyzed and matched with a function identified in a hierarchy of elements. The hierarchy is applicable to several mobile apps. A category of the function identified in the hierarchy is assigned to the selected element. The category and a numerosity of occurrences of the selected element in the mobile app are output in a blueprint of the mobile app.

Creating an on-demand blueprint of a mobile application

Growing community of inventors

Western Springs, IL, United States of America

Blaine H Dolph