International Business Machines Corporation

Gary William Reiss

Aaron K Baughman

Shikhar Kwatra

Gray Franklin Cannon

Qiqing Christine Ouyang

Stephen Carl Hammer

Corey B Shelton

Nancy Anne Greco

David Alvra Wood, Iii

Eduardo Morales

Sara Perelman

Michael Choong

Kevin Lee Masters

Edge function-guided artificial intelligence (AI) request routing is provided by applying a machine learning model to predictors of cloud endpoint hydration to determine hydration levels of cloud endpoints, of a hybrid cloud environment, that provide AI processing, determining, for each edge component of a plurality of edge components of the hybrid cloud environment and each cloud endpoint of the cloud endpoints, alternative flow paths between the edge component and the cloud endpoint, the alternative flow paths being differing routes for routing data between the edge component and the cloud endpoint, and the alternative flow paths being of varying flow rates determined based on the hydration levels of the cloud endpoints, and dynamically deploying edge functions on edge component(s), the edge functions configuring the edge component(s) to alternate among the alternative flow paths available in routing AI processing requests from the edge component(s) to target cloud endpoints of the cloud endpoints.

Edge function-guided artificial intelligence request routing

A method, computer system, and a computer program product for speech synthesis is provided. The present invention may include generating one or more final voiceprints. The present invention may include generating one or more voice clones based on the one or more final voiceprints. The present invention may include classifying the one or more voice clones into a grouping using a language model, wherein the language model is trained using manually classified uncloned voice samples. The present invention may include identifying a cluster within the grouping, wherein the cluster is identified by determining a difference between corresponding vectors of the one or more voice clones below a similarity threshold. The present invention may include generating a new archetypal voice by blending the one or more voice clones of the cluster where the difference between the corresponding vectors is below the similarity threshold.

Voice cloning transfer for speech synthesis

According to a first aspect of the present invention, there is provided a computer implemented method, a computer system and a computer program product, including training a set of exploitation models, training a set of exploration models, generating a combined exploitation and exploration heat map, and inputting the combined exploitation and exploration heat map into a convoluted neural network.

Deep evolved strategies with reinforcement

An approach for altering alter training data and training process associated with a neural network to emulate environmental noise and operational instrument error by using the concepts of shots to sample within a squeezed space model, wherein shots are an uncertainty index that is the average of all shots from a sampling, is disclosed. The approach leverages a squeeze theorem to create a squeezed space model based on the regression of the upper and lower bound associated with the environmental noise and instrument error. The approach calculates an average noise index based on the squeezed space model, wherein the index is used to alter the training data and process.

Modeling environment noise for training neural networks

A computer-implemented method includes: receiving, by a computer device, sensor data for a plurality of UAVs in a fleet of UAVs; applying, by the computer device, logistic regression to the sensor data; predicting, by the computer device, a probability of malfunction of each UAV in the fleet of UAVs based on the applying; combining, by the computer device, the probability of malfunction of each UAV with a pre-existing malfunction data set to produce an intermediate malfunction data set; generating, by the computer device, additional cases of predicted UAV malfunctions with a GAN, the GAN using the intermediate malfunction data set as initial training data for the GAN; combining, by the computer device, the additional cases with the intermediate malfunction data set to produce a combined malfunction data set; and comparing, by the computer device, the sensor data for a first UAV of the UAVs to the combined malfunction data set.

Unmanned aerial vehicle maintenance and utility plan

A method, a computer system, and a computer program product for parallel conversion is provided. Embodiments of the present invention may include analyzing raw classical code using a code embedded deep learning model. Embodiments of the present invention may include analyzing running classical code using a deep learning model. Embodiments of the present invention may include marking a location of the raw classical code for a first quantum conversion. Embodiments of the present invention may include suggesting a memory size of the running classical code for a second quantum conversion. Embodiments of the present invention may include aggregating the raw classical code for the first quantum conversion. Embodiments of the present invention may include aggregating the running classical code for the second quantum conversion.

Classical artificial intelligence (AI) and probability based code infusion

A method for voice signal fatigue compensation, that includes sampling, using an audio signal capturing apparatus, a segment of a voice signal in a normal time series to form a normal series sample, generating, using a processor and a memory, from the normal series sample, a reversed series sample, and constructing, by executing using the processor and the memory a time-series mixing component, a first synthesized segment by mixing the normal series sample and the reversed series sample, the first synthesized segment including a compensation for an instance of micro fatigue in the segment of the voice signal. The method also includes forming a fatigue-compensated voice segment from the first synthesized segment, and outputting, as a fatigue-compensated voice segment, the first synthesized segment.

Trending and context fatigue compensation in a voice signal

An audio forecasting algorithm that is adjusted (or trained), by machine learning, prior to a sports contest that will be broadcast. The audio forecasting algorithm is then used to position a set of mobile microphones on an ongoing basis during the sports contest. In some embodiments, a band forecasting algorithm is used in the audio forecasting algorithm. In some embodiments, a swarm based correlation algorithm is used in the audio forecasting algorithm.

Growing community of inventors

Buford, GA, United States of America

Gary William Reiss