Abstract: Targeted robotic process automation (RPA) bot decisioning includes building a user-specific decisioning profile of a user, where the user-specific decisioning profile characterizing user decision-making in performing task actions as part of workflow execution on a user device, monitoring for deployment of workflow tasks and identifying that a deployed workflow task includes a decisioned task action, selecting the user-specific decisioning profile, and executing an RPA bot to perform the decisioned task action based on the user-specific decisioning profile, in which decisioning criteria informed by the user-specific decisioning profile serves as a basis for the execution of the RPA bot in automatically performing the decisioned task action.

Abstract: A method, by an intelligent access control system, of implementing access control to a controlled area having a plurality access areas divided by at least one access point includes the following operations. A security token is generated. The security token is transmitted to a client module executing on a client device associated with a first user, and the client module is caused to pair with an access control device associated with the first user. The client module is caused to transfer the security token to the access control device. A physical location of the client device within the controlled area is monitored. A determination is made, by an access control system and based upon the physical location, that the first user is to be revalidated. The client module causes, based upon the determining, the client device to generate an alert.

Abstract: Segmentation of personalized content from media content of events includes retrieving media content associated with an event. In data including a set of attributes associated with a user is retrieved. The user is one of an active participant of the event or a passive participant of the event. A machine learning (ML) model is applied to the input data and the media content. Additionally, a segment of the media content correlated with the set of attributes is determined. At least one an alert or the segment is rendered on a user device associated with the user.

Abstract: Various embodiments are provided herein for making ergonomic keybinding adjustments based on user behavioral inputs in a computing environment having one or more processor devices. One or more metrics of a user input are measured. If the one or more metrics are determined to exceed a predetermined threshold, one or more of the user behavioral inputs associated with the one or more metrics are identified to be provided to a keybinding adjustment o the one or more behavioral inputs.

Abstract: Mechanisms are provided to automatically identify referenced assets in real-time streaming communications and automatically identify, retrieve, and present these assets during the real-time streaming communication. The mechanisms obtain, in real-time, a textual representation of a user's input during the real-time streaming communication. As part of a background process to the real-time streaming communication, computer natural language processing is executed on the textual representation to identify key terms or key phrases from the textual representation. The computer natural language processing is specifically configured to identify key terms/phrases corresponding to references to assets. As part of the background process, one or more search queries are generated and executed to search for a matching asset based on the identified key terms/phrases. As part of the real-time streaming communication, the matching asset is presented to participant computing devices of the real-time streaming communication.

Abstract: Managing opt-in audio input is provided. A set of insights corresponding to an upcoming in-person event is generated using a set of machine learning models based on keyword frequency, predicted event attendee sentiment, predicted most discussed event topics, predicted frequently asked event questions, and predicted event outcome. An event summary for the upcoming in-person event is generated using the set of insights corresponding to the upcoming in-person event that was generated based on the keyword frequency, the predicted event attendee sentiment, the predicted most discussed event topics, the predicted frequently asked event questions, and the predicted event outcome. The event summary for the upcoming in-person event is distributed using a distribution channel to an event organizer and a plurality of event attendees who will be physically present at the in-person event.

Abstract: A computer-implemented method for training a large language model (LLM) on a plurality of programming languages to recognize common coding patterns and generating, using the trained LLM, software code based on user-defined inputs. The method further includes analyzing the generated software code to identify destructive code and determining a reversal of the generated software code from a current state to a previous state is required based on the destructive code. Responsive to the determining that a reversal is required, the method may include running the generated software code comprising the destructive code in a controlled and isolated environment and providing a user with a reversal function that causes the generated software code to revert from the current state to the previous state. The method further includes executing the generated software code.

Abstract: Aspects of the present disclosure relate to virtual diagram generation and management. Speech data can be obtained from a speaker within a web conference. A set of key phrases can be extracted from the speech data within the web conference based on a first set of user data associated with a first user. A virtual diagram can be generated using the set of key phrases, wherein the virtual diagram is generated based on the first set of user data. The virtual diagram can be caused to be displayed to the first user within the web conference.

Abstract: A computer-implemented method includes capturing, using one or more cameras, user interactions with a printed three-dimensional (3D) object. The printed 3D object has a digital twin. The user interactions, captured by the one or more cameras, are interpreted using the digital twin. Actions are generated based on the interpreted user interactions.

Abstract: In an omnichannel call center, a voice call between a caller and an agent is monitored, including generating text of the voice call in real time, and using Natural language processing (NLP) on the text to determine whether the call has been prematurely disconnected. If the call was prematurely disconnected, an alternative communication channel between the caller and the agent is established.

Abstract: A computer-implemented method, a computer system and a computer program product customize delivery of transactional documents. The method includes receiving a transactional document associated with a user. The method also includes obtaining a preferred document format for the user and a preferred delivery method for the user from a database. The method further includes identifying a current display format of the transactional document. In addition, the method includes modifying the current display format of the transactional document, wherein a modified transactional document has the preferred document format for the user. Lastly, the method includes transmitting the modified transactional document to the user using the preferred delivery method for the user.

Abstract: A computer-implemented method for training a large language model (LLM) on a plurality of programming languages to recognize common coding patterns and generating, using the trained LLM, software code based on user-defined inputs. The method further includes analyzing the generated software code to identify destructive code and determining a reversal of the generated software code from a current state to a previous state is required based on the destructive code. Responsive to the determining that a reversal is required, the method may include running the generated software code comprising the destructive code in a controlled and isolated environment and providing a user with a reversal function that causes the generated software code to revert from the current state to the previous state. The method further includes executing the generated software code.

Abstract: Generation and use of screenshots of interactive displays are provided. A screenshot is received. The screenshot comprises an area of an interactive display. An interactive object is identified within the area of the interactive display. The interactive object corresponds to a function. A simulated interactive object corresponding to the interactive object is generated. The simulated interactive object corresponds to the function. A composite screenshot comprising the screenshot and the simulated interactive object is rendered for display, on a display.

Abstract: Embodiments of the disclosure provide systems and methods for implementing 3D printing to manufacture 3D print objects with dynamic adjustment of 3D print infill specifications based on determined regions of stress points in a given 3D device design. A disclosed method comprises receiving a 3D object design for a 3D device to be printed; determining, based on the 3D object design and a trained model, a region of the 3D object with a projected likelihood of human interaction above a threshold likelihood; and generating, based on the determined region of the 3D object, instructions to selectively adjust one or more print infill specifications at the region.

Abstract: An embodiment includes detecting by a Hardware Management Component of a datacenter, a failure of a hardware in the datacenter. The embodiment includes responsive to the detecting, determining by an Asset Manager Component, a lead time for a replacement of the failed hardware. The embodiment includes determining by the Hardware Management Component a replacement decision based on a priority of a workload deployed on the failed hardware and the lead time. The embodiment includes deciding by the Hardware Management Component whether to replace the failed hardware with a substitute hardware from a lower priority system of the datacenter based on the replacement decision. The embodiment also includes adjusting by a Workload Controller Component the workload deployed on the failed hardware based on the replacement decision.

Abstract: Techniques are described with respect to a system, method, and computer program product for merging augmented reality experiences into an augmented reality environment. An associated method includes determining a first augmented reality experience of a first user and a second augmented reality experience of a second user; analyzing a plurality of environmental contextual data associated with the first augmented reality experience and the second augmented reality experience; and based on the analysis, generating a merged augmented reality environment comprising the first user and the second user.

Abstract: Embodiments of the present invention provide computer-implemented methods, computer program product, and computer systems. One or more processors can monitor file revision history of a filesystem. The one or more processors generate a remedial action based on previous user interaction behavior in response to detecting a failed interaction with a file in the filesystem. The one or more processors execute the remedial action that retrieves the file.

Abstract: Methods and systems for heat treatment include determining heat treatment needs for a three-dimensional (3D) object. A heating structure is designed to supply the heat treatment needs of the 3D object. The heating structure is assembled from modular heating structures that include heating elements. Heat treatment is performed on the 3D object using the heating structure.

Abstract: Dynamic updating of content viewership alignment based on aggregate viewer metadata is provided. In some embodiments, a virtual meeting schedule is read. The virtual meeting schedule includes a list of attendees and one or more pieces of content. Attendee data and content history are captured for each attendee based on the list of attendees. The content history is aggregated for the list of attendees. A relevancy score is generated for each of the one or more pieces of content based on the aggregated content history. Generating the relevancy score comprises determining a similarity component between each piece of content and the content history, determining a familiarity component based on each attendee, and an interaction component based on the content history. Each piece of content is annotated according to its relevancy score. The annotated one or more pieces of content are outputted in the virtual presentation system.

Abstract: According to one embodiment, a method, computer system, and computer program product for providing optimized automated note-taking. The embodiment may include transcribing audio data of a current user interaction into text data. The embodiment may include analyzing the text data using natural language processing to identify a context of the current user interaction. The embodiment may include generating, via a generative artificial intelligence (GenAI) model, one or more summaries of contextually relevant portions of the text data. The embodiment may include creating one or more prompts for the GenAI model based on the one or more summaries. The embodiment may include querying the GenAI model using the one or more prompts and using one or more existing prompts from a previous user interaction to identify a contextual opportunity for notification. The embodiment may include notifying a user of the contextual opportunity for notification during the current user interaction.