International Business Machines Corporation

Kyndryl, Inc.

Gary Francis Diamanti

Mauro Marzorati

Aaron K Baughman

Jeremy R Fox

Iwao Hatanaka

Sarbajit Kumar Rakshit

Elizabeth Merckel Valletti

Craig M Trim

Stephen Joseph Kenna

Patrick Rodrigo Mardones Rodriguez

Shikhar Kwatra

Moncef Benboubakeur

William A Mills

Raghuram Srinivasan

Martin G Keen

Zachary A Silverstein

Jana Helton Jenkins

Diwesh Pandey

Christian Eggenberger

Ryan R Anderson

Yaser K Doleh

Julija Narodicka

Seda Özses

Rodrigo Goulart Silva

John Christian Fluke

Nicholas Michael Wilkin

Jorge Alonso Hernandez Nevarez

Jeffrey W Lucas

Juraj Nyiri

Marco Antonio Martinez Ruvalcaba

Ann Bieghler Shirk

A computer-implemented process for training a natural language processing (NLP) agent having a reinforced learning model includes the following operations. A type of document from a document corpus is identified using metadata particularly associated with the document. The NLP agent tokenizes the document to generate a plurality of tokens. Using a schema identified from the type of the document, one of the plurality of tokens is compared to a system of record (SOR) field from the schema. A similarity score between the one of the plurality of tokens with a correct value and a reward based upon the similarity score are generated. A determination is made that an optimum minimum average similarity rate has not been obtained. Based upon the determination, the reinforced learning model is trained using a loss function that includes the reward.

Machine learning for training NLP agent

Methods, computer program products, and systems are presented. The methods include, for instance: building, by one or more DNA processor, a DNA strand corresponding to a conditional expectation. The methods include, for instance: obtaining, by one or more DNA processor, a conditional expectation having a regularization metric.

Machine learning (ML) modeling by DNA computing

A mechanism is provided in a data processing system for question answering using time weighted evidence. The mechanism receives an input question. The mechanism determines a time focus for the input question and defines a weighting function. The weighting function is a bell curve having a peak at the time focus on a time axis. The mechanism decomposes the input question into one or more queries and applies the one or more queries to a corpus of information to obtain a set of hypothesis evidence. Each item of information within the hypothesis evidence has an associated time value. The mechanism weights the set of hypothesis evidence based on the associated time values according to the weighting function to form time weighted evidence and generates hypotheses for answering the input question based on the time weighted evidence.

Question answering with time-based weighting

A computer receives a multimedia data, where the multimedia data comprises a plurality of frames. The computer converts the multimedia data into a signal wave having a plurality of frequencies and a plurality of amplitudes. The computer determines a frame from the plurality of frames having a pronoun. The computer identifies a topic of the frame. The computer searches for a frame in a media repository having a highest correlation coefficient with the topic of the frame, where the frame from the media repository comprises a bag of objects and resolves the anaphora disambiguation by substituting the pronoun with an object from the bag of objects.

Anaphora resolution for enhanced context switching

A method and a system for dataset evaluation are provided. The method can include acquiring a new dataset for integration with an existing dataset. The existing dataset is used to train a machine learning model. The method can also include calculating a baseline of variation for the existing dataset. The baseline of variation can be determined by evaluating various characteristics of the existing dataset. The method can further include determining an output value for the new dataset by also evaluating characteristics of the new dataset. The method can also include comparing the output value to the baseline of variation to generate a variance between the output value and the baseline of variation. The method can further include determining whether the variance is within an acceptable range of the baseline of variation.

Detecting and preventing unwanted model training data

A computer-implemented method reduces toil in chat systems. The method includes receiving, from a sending account, a message to be sent in a chat environment, where the chat environment includes a plurality of accounts including the sending account. The method also includes determining, by a learning model, a category of the message. The method further includes generating, for each account of the plurality accounts in a channel of the chat environment, a score, where the score indicates a likelihood an associated account will respond to a post with relevant information. The method includes recommending, based on the score, one or more accounts of the plurality of accounts to an at message, where the at message is configured to cause a notification to the associated account. The method further includes posting the message in the channel of the chat environment while at-mentioning the one or more accounts.

Reducing toil in chat environments

An embodiment for creating a template for user experience by segmenting a visual surrounding is provided. The embodiment may include receiving real-time and historical data relating to one or more content interactions of a user wearing an augmented reality (AR) device. The embodiment may also include analyzing one or more contextual situations of the one or more content interactions. The embodiment may further include identifying one or more objects of interest in a visual surrounding environment of the user. The embodiment may also include in response to determining the identification of the object type is confident, predicting a contextual need for each object of interest. The embodiment may further include creating one or more information display templates. The embodiment may also include populating the one or more information display templates with information and displaying the one or more populated information display templates to the user.

Segmenting visual surrounding to create template for user experience

A wearable computing device generates a volumetric visualization at a first position that is located in a three-dimensional space. The wearable computing device includes a volumetric source configured to create the volumetric visualization. The wearable computing device includes one or more sensors configured to determine movement of the wearable computing device. A movement of the wearable computing device is identified by the wearable computing device. Based on the movement the wearable computing device adjusts the volumetric source.

Mid-air volumetric visualization movement compensation

Methods, computer program products, and systems are presented. The methods include, for instance: identifying a training data set and defining a window for an initial beta value representing bias tolerated in formulating expectation conditional to each feature vector from the training data set. The conditional expectations are parallelly regularized by use of DNA computer. Amongst numerous combinations of candidate models, a best fit ensemble is produced as the machine learning model for predicting targeted outcomes based on inputs other than the training data set.

In an approach for building a machine learning model that predicts the appropriate action to resolve a malfunction or system error, a processor receives an alert that a malfunction or a system error has occurred. A processor creates a workspace on a ChatOps platform integrated with a chatbot and one or more tools. A processor inputs data relating to the alert in a natural language format. A processor processes the data using a natural language processing algorithm. Responsive to determining a pre-set threshold for outputting the appropriate action is not met, a processor establishes a conversation between two or more support service agents in the workspace. A processor monitors the conversation using the natural language processing algorithm. A processor analyzes a transcript of the conversation using text analytics or pattern matching. A processor creates and trains a machine learning model to predict the appropriate action in future iterations.

System and AI pattern model for actionable alerts for events within a ChatOps platform

Visualizing a contaminant is provided. A contaminant of a plurality of different contaminants included in a contaminant knowledgebase is identified based on analysis of contaminant-relevant data received from one or more sensors of a plurality of different sensor arrays regarding an enclosed physical space. A concentration and a type of the contaminant is identified based on the contaminant-relevant data and information included in the contaminant knowledgebase. A location of the contaminant is identified within the enclosed physical space based on location of the one or more sensors that obtained the contaminant-relevant data and a digital twin of the enclosed physical space. A visualization of the contaminant is projected at an area proximate to the location of the contaminant using a holographic image indicating the concentration and the type of the contaminant within the enclosed physical space.

Generating a holographic image to visualize contaminants

Aspects of the present invention provide an approach for non-visually mapping an environment. In an embodiment, a set of images that is within the field of view of the user is captured from a video capture device worn by the user. Environmental text that is within the set of images is identified. An aural contextual indicator that corresponds to the environmental text is then generated. This aural contextual indicator indicates the informational nature of the environmental text. An aural map of the environment is created using a sequence of the generated aural contextual indicators. This aural map is delivered to the user via an aural output device worn by the user in response to a user request.

Non-visual environment mapping

A method and system for improving real-time audio, video, and sensor detection and analysis is provided. The method includes retrieving audio/video data, biometric data, and environmental data associated with associated with a user at a location. The audio/video data, biometric data, and environmental data are analyzed and based on the analysis, a current biometric state of the user is determined. The current biometric state of the user is compared to a baseline biometric state of the user and it is determined that the current biometric state comprises an elevated biometric state with respect to the baseline biometric state of the user. Self-learning software code for executing a machine based interaction modification event associated with reducing the elevated biometric state of the user is generated and the machine based interaction modification event is executed.

Real time sensor attribute detection and analysis

Generating a natural language text response to a user query is provided. Sentiment attributes of the user query are encoded into qubits as particle spins. The particle spins in encoded sentiment qubits are observed to generate spin data. A sentiment value is derived for each sentiment key term of the user query based on the generated spin data. A total sentiment of the user query is determined by combining derived sentiment values corresponding to each sentiment key term of the user query. A natural language text response to the user query is generated based on the determined total sentiment of the user query. The natural language text response is transmitted to a client device via a network.

Quantum superposition and entanglement of social sentiment and natural language generation

Systems and methods for multi-factor authentication of a caller identification (ID) identifier are disclosed. In embodiments, a method includes: receiving, by a computing device of a call recipient, a call request message from a remote caller device of a caller, the call request message including an authentication token shared between the caller and the call recipient, and an identifier; authenticating the caller using the authentication token; saving an authorized caller record in an authorized caller list on the computing device, the authorized caller record including the identifier; receiving an incoming telephone call from the caller, the incoming call including a caller ID identifier; determining that the caller ID identifier matches the identifier in the authorized caller record; and initiating completion of the call based on the determining that the caller ID identifier matches the identifier in the authorized caller record.

Growing community of inventors

Wake Forest, NC, United States of America

Gary Francis Diamanti