Abstract: A supervised discriminator for detecting bio-markers in an audio sample dataset is trained and a denoising autoencoder is trained to learn a latent space that is used to reconstruct an output audio sample with a same fidelity as an input audio sample of the audio sample dataset. A conditional auxiliary generative adversarial network (GAN) trained to generate the output audio sample with the same fidelity as the input audio sample, wherein the output audio sample is void of the bio-markers. The conditional auxiliary generative adversarial network (GAN), the corresponding supervised discriminator, and the corresponding denoising autoencoder are deployed in an audio processing system.

Abstract: Methods, systems, and computer program products for auto-discovery of reasoning knowledge graphs in supply chains are provided herein. A computer-implemented method includes obtaining a spatiotemporal query related to a demand of at least one product in a supply chain; analyzing the spatiotemporal query to identify one or more parameters affecting the demand of the at least one product, wherein the one or more parameters comprise at least one of one or more climate parameters and one or more disruptive event parameters; generating a knowledge graph comprising information indicating an impact on the demand of the at least one product for at least a portion of the one or more parameters; and outputting, to a user interface, an explanation of a predicted demand forecast for the at least one product based at least in part on the knowledge graph.

Abstract: Concepts are proposed for augmenting a plurality of climate impact and hazard models. In particular, the plurality of climate impact and hazard models are modified/altered based on a user-specified requirement (i.e., a specified geographical location, temporal location or hazard type) and at least one intelligent (i.e., AI-enabled) workflow. Each modified climate impact and hazard model is then analyzed/processed using a machine learning model to identify model parameters which may improve the climate impact and machine learning model. An operation is then identified and executed based on the intelligent workflow, and model parameter. In this way, speed, efficiency and accuracy of climate impact and hazard models may be improved at scale.

Abstract: A method, computer system, and a computer program product for interpreting conference call interruptions is provided. The present invention may include analyzing a conference call including at least two participants. The present invention may also include identifying at least one interrupted segment in the analyzed conference call. The present invention may further include reconstructing the identified at least one interrupted segment of the analyzed conference call and integrating it with the uninterrupted (heard) portion of the conference call.

Abstract: Train a machine learning model, using an image-based knowledge graph of tropical cyclone data, for implementing a surface field modeling architecture that produces images of at least surface wind fields and surface rainfall fields from images of at least tropical cyclone tracks and pressure intensities. Generate model images of a modeled surface wind field and a modeled surface rainfall field by providing images of at least a user-generated tropical cyclone track and pressure intensity to the trained machine learning model.

Abstract: An exemplary method includes vehicle-mounted sensors continuously detecting vehicle speed and vehicle tire and steering vibrations; a processor implementing a machine-learning program that continuously monitors signals from the vehicle-mounted sensors and compares detected vehicle tire and steering vibrations to upper bounds corresponding to detected vehicle speed; and the processor alerting a vehicle driver that wheel or tire service is required based on the detected vehicle tire and steering vibrations exceeding the upper bounds. An exemplary apparatus includes a vehicle; tires mounted to the vehicle; a speed sensor mounted to the vehicle; a vibration sensor mounted to the vehicle; and a processor connected in communication with the speed sensor and the vibration sensor. The processor is adapted to implement any of the method steps above.

Abstract: Techniques are described for providing live first aid response guidance using a machine learning based cognitive aid planner. In one embodiment, a computer-implemented method is provided that comprises classifying, by a system operatively coupled to a processor, a type of an injury endured by a patient. The method further comprises, employing, by the system, one or more machine learning models to estimate a risk level associated with the injury based on the type of the injury and a current context of the patient.

Abstract: One or more processors receive data associated with delivery of a package to a delivery drop-off area. The data includes delivery history of the drop-off area, current and predicted weather conditions, customer profile data, and package content information. The one or more processors detect surface conditions of the delivery drop-off area. The one or more processors determine a hazardous condition for retrieval of a delivered package in the delivery drop-off area, based on the images and sensor data received from the delivery vehicle and processed by one or more machine learning models. The one or more processors perform actions to mitigate the hazardous condition at the delivery drop-off area, and the one or more processors provide a notification updating the hazardous condition in the delivery drop-off area and the at least one mitigating action taken.

Abstract: Embodiments of the present invention provide a computer system, a computer program product, and a method that comprises identifying a plurality of code datasets prior to a data migration; analyzing the identified code datasets for a plurality of parameters; dynamically predicting a carbon footprint associated with the analyzed code datasets based on the plurality of parameters for each analyzed code dataset; and automatically optimizing the analyzed code datasets based on the predicted carbon footprint for data migration.

Abstract: A method comprising retrieving a plurality of invalid user first commands and a plurality of user responses stored in a memory, wherein each of plurality of invalid user verbal commands are commands to a smart hub requesting a first smart device perform an action, wherein each of the plurality of invalid commands includes at least a first name for the first smart device, wherein the first name is not a valid name for the first smart device. Determining a trend within the retrieved plurality of invalid user verbal commands and a plurality of user responses, wherein the trend identifies the first name as being used a plurality of times and identifies the first smart device the user was referring to when the user used the first name. Automatically updating a namespace database to include the first name as an alias for a setup name for the first smart device.

Abstract: Aspects of the present invention disclose a method for synchronizing a voice reply from AI voice assistant based on user activities. The method includes one or more processors identifying a task of a user that corresponds to a voice query of the user. The method further includes generating a sequence of sub-activities of the task corresponding to the voice query of the user. The method further includes determining a completion status of each sub-activity of the sequence of sub-activities of the task corresponding to the voice query of the user based at least in part on activity data received from one or more computing devices in an operating environment of the user. The method further includes synchronizing a voice reply of a computing device with the sequence of sub-activities of the task based at least in part on the completion status of each sub-activity of the sequence.

Abstract: An illustrative embodiment includes a method for detecting whether a given item within input data is anomalous. The method includes: extracting activations at respective nodes of an autoencoder trained on the data, the activations comprising activations at the respective nodes for the given item within the data and for one or more other items within the data; calculating p-values corresponding to the respective nodes, wherein the p-value for a given node represents a proportion of the activations at the given node for the other items which are greater than the activations at the given node for the given item; determining at least one score at least in part by scanning for a subset of the respective nodes which maximizes a statistical scoring function applied to the corresponding p-values; and providing a visualization of at least the given item within the input data to a human user.

Abstract: Embodiments of the present invention describe predictively reconstructing a physical event using augmented reality. Embodiments describe, identifying relative states of objects located in a physical event area by using video analysis to analyze collected video feeds from the physical event area before and after a physical event involving at least one of the objects, creating a knowledge corpus including the video analysis and the collected video feeds associated with the physical event and historical information, and capturing data, by a computing device, of the physical event area.

Abstract: User interactions with a supply chain system are monitored based on a tracked ontology enrichment process, an explainable reasoning graph is constructed based on the monitored user interactions and domain specific reasoning information; and an explainable insight of the monitored user interactions is learned, as is a user interaction embedding for an embedding space, based on the constructed explainable reasoning graph and the explainable insight. External data is incorporated into the embedding space, a joint embedding is learned based on the user interaction embedding, and missing entities and relationships are identified for incorporation into an ontology based on the user interactions and joint embedding. The ontology is revised to incorporate the missing entities and relationships into the ontology to create a revised ontology, and a supply chain is controlled based on the revised ontology.

Abstract: Spatio-temporal climate forecasts are analyzed and one or more resiliency policies for a supply chain are dynamically generated. The resiliency policy is embedded in a resiliency reasoning graph and a temporal feedback loop is performed based on user feedback regarding the generated resiliency policy and user interaction with the resiliency reasoning graph. One or more machine learning models are updated based on the user feedback and a joint optimization of the machine learning models is re-solved based on the user feedback. The resiliency policy is updated based on the updated machine learning models based on the user feedback and an operation of a supply chain is adjusted based on the updated resiliency policy.

Abstract: Initial and boundary conditions, and parameters associated with geophysical modeling can be received. Based on the received initial and boundary conditions and parameters, a multiscale model can be trained for data generation to produce first resolution simulation data and second resolution simulation data for a surrogate machine learning model training, where the second resolution simulation data has higher resolution than the first resolution simulation data. A surrogate model can be created using neural operators, where the surrogate model is trained using the first resolution simulation data and second resolution simulation data. An operational forecasting model can be generated using the surrogate model.

Abstract: A method and system generate news headlines from user input parameters. The user input parameters include a specified geographic region of interest and an industry of interest. Climate data and carbon emissions data for the specified geographic region of interest is retrieved. Supply chain dependencies are determined. A machine learning model is generated using the specified geographic region of interest, the industry, the climate data, the carbon emissions data, and the supply chain dependencies. The machine learning model performs an impact analysis on a supply chain based on the climate data and the carbon emissions data. The machine learning model predicts a supply chain performance for the industry based on the impact analysis. A news headline is automatically generated describing the predicted supply chain performance. The news headline includes an underlying basis for the predicted supply chain performance.

Abstract: A first set of user data is received and a user profile is constructed based on the user data and in accordance with a sensitive service involving the user. A situational context is analyzed based on the first set of data. Personalized questions are generated, responsive to the user profile and to the situational context. The personalized questions are presented to a user corresponding to the user data and responses to same are received, including detection of user micro-expressions. The responses are analyzed, according to one or more machine learning models. A neural network model selects an action to be performed in response to analyzing the responses from the user; the action is a sensitive service involving the user. An apparatus is triggered to send a simple message service (SMS) message to a point of care service professional; the message recommends performance of the sensitive service on the user.

Abstract: Methods, systems, and computer program products for dynamically enhancing supply chain strategies based on carbon emission targets are provided herein.

Abstract: Predicting infection risk by generating a first temporal graph of a first set of disease progression data, generating a second temporal graph of a second set of disease progression data, combining a first temporal graph node embedding and a second temporal graph node embedding, and generating a predicted infection risk according to the first temporal graph node embedding and the second temporal graph node embedding.