International Business Machines Corporation

Nuance Communications, Inc.

Utopus Insights, Inc.

Kyndryl, Inc.

Universiti Brunei Darussalam

Ashish Verma

Sachindra Joshi

Shantanu Ravindra Godbole

Om Dadaji Deshmukh

Nitendra Rajput

Indrajit Bhattacharya

Kevin W English

Jeffrey Mark Achtermann

Ashwin Srinivasan

Yogish Sabharwal

Jitendra Ajmera

Prasanta Kumar Ghosh

Chalapathy Venkata Neti

Srujana Merugu

L Venkata Subramaniam

Shajith I Mohamed

Saurav Basu

Karthik Visweswariah

Sankar Basu

Jayaram Kallapalayam Radhakrishnan

Vaibhav Saxena

Anamitra Roy Choudhury

Saurabh Goyal

Himabindu Lakkaraju

Bernard Andrew Roden

Saket Saurabh

Suraj Satishkumar Sheth

Manish Anand Bhide

Andrew William Senior

Parijat Dube

Sai Zeng

Ravi Kothari

Ajay Gupta

Chitra Dorai

Jagabondhu Hazra

Etienne Marcheret

Noam Slonim

Sumit Mittal

Danish Contractor

Manikandan Padmanaban

Sreeram Viswanath Balakrishnan

Dipanjan Chakraborty

Amrita Saha

Tanveer Afzal Faruquie

Kenneth W Church

Vikas Raykar

Jeffrey N Marcus

Koyel Mukherjee

Vikas Chandan

Chumki Basu

Sunil Chandra

Pankaj Kankar

Thomas George

Venkatesan T Chakaravarthy

Shrihari Vasudevan

Rashmi Mittal

Abhinanda Sarkar

Jayadeva

Sathyajith Mathew

Sandeep Jindal

Mitesh Khapra

Saurabh Manish Raje

Shailesh Joshi

Chandrasekar Radhakrishnan

Pooja Kumar

Sudha Krishnamurthy

Arpita Ghosh

Sidharth Gupta

Tanveer Atzal Faruquie

Alind Khare

Jeffrey M Achtermann

Shantanu R Godbole

Katyaini H Naga

Kartik Audhkhasi

Kundan Kandhway

Maureen E Rzasa

Mohammad Iskandar Pg Hj Petra

Gary J Wright

Himanshu Pant

Rajni Bajaj

Mridula Bhandari

Lyndon J D'Silva

Methods, systems, and computer program products for cognitive disambiguation of problem-solving tasks involving a power grid are provided herein. A computer-implemented method includes capturing user feedback pertaining to relevance of remote terminal unit measurements related to a grid event through user interface interactions carried out by the user, wherein the user interface is communicatively linked to at least one computing device; automatically inferring rules related to the grid event to curate remote terminal unit measurements across iterations of analysis by recognizing irrelevant data and/or distractions in a visual display associated with the user interface, wherein said automatically inferring comprises implementing machine learning via the at least one computing device based on the user feedback; and outputting candidate solutions to a problem-solving task involving the grid based on the inferred rules, wherein said outputting is carried out by the at least one computing device.

Cognitive disambiguation for problem-solving tasks involving a power grid using relevance feedback

A method and system. Target clusterability is calculated as an average of a respective clusterability of at least one target data item comprised by a target domain. Target-side matchability is calculated as an average of a respective matchability of each target centroid of the target domain to source centroids of a source domain, wherein the source domain comprises at least one source data item. Source-side matchability is calculated as an average of a respective matchability of each source centroid of said source centroids to the target centroids. Source-target pair matchability is calculated as an average of the target-side matchability and the source-side matchability. Cross-domain clusterability between the target domain and the source domain is calculated as a linear combination of the calculated target clusterability and the calculated source-target pair matchability. The cross-domain clusterability is transferred to a device.

Cross-domain clusterability evaluation for cross-guided data clustering based on alignment between data domains

Methods, systems, and computer program products for accelerating inference of transformer-based models are provided herein. A computer-implemented method includes obtaining a machine learning model comprising a plurality of transformer blocks, a task, and a natural language dataset; generating a compressed version of the machine learning model based on the task and the natural language dataset, wherein the generating comprises: obtaining at least one set of tokens, wherein each token in the set corresponds to one of the items in the natural language dataset, identifying and removing one or more redundant output activations of different ones of the plurality of transformer blocks for the at least one set of tokens, and adding one or more input activations corresponding to the one or more removed output activations into the machine learning model at subsequent ones of the plurality of the transformer blocks; and outputting the compressed version of the machine learning model to at least one user.

Accelerating inference of transformer-based models

One embodiment provides a method, including: receiving at least one deep learning job for scheduling and running on a distributed system comprising a plurality of nodes; receiving a batch size range indicating a minimum batch size and a maximum batch size that can be utilized for running the at least one deep learning job; determining a plurality of runtime estimations for running the at least one deep learning job; creating a list of optimal combinations of (i) batch sizes and (ii) numbers of the plurality of nodes for running both (a) the at least one deep learning job and (b) current deep learning jobs; and scheduling the at least one deep-learning job at the distributed system, responsive to identifying, by utilizing the list, that the distributed system has necessary processing resources for running both (iii) the at least one deep learning job and (iv) the current deep learning jobs.

Application aware resource allocation for deep learning job scheduling

A computer-implemented machine learning model training method and resulting machine learning model. One embodiment of the method may comprise receiving at a computer memory training data; and training on a computer processor a machine learning model on the received training data using a plurality of batch sizes to produce a trained processor. The training may include calculating a plurality of activations during a forward pass of the training and discarding at least some of the calculated plurality of activations after the forward pass of the training.

Model training with variable batch sizing and gradient checkpoint segments

Systems, computer-implemented methods, and computer program products that can facilitate switching a model training process from a ground truth training phase to an adversarial training phase based on performance of a model trained in the ground truth training phase are provided. According to an embodiment, a system can comprise a memory that stores computer executable components and a processor that executes the computer executable components stored in the memory. The computer executable components can comprise an analysis component that identifies a performance condition of a model trained in a model training process. The computer executable components can further comprise a trainer component that switches the model training process from a ground truth training process to an adversarial training process based on the identified performance condition.

Performance based switching of a model training process

Methods, systems, and computer program products for determining operating range of hyperparameters are provided herein. A computer-implemented method includes obtaining a machine learning model, a list of candidate values for a hyperparameter, and a dataset; performing one or more hyperparameter range trials based on the machine learning model, the list of candidate values for the hyperparameter, and the dataset; automatically determining an operating range of the hyperparameter based on the one or more hyperparameter range trials; and training the machine learning model to convergence based at least in part on the determined operating range.

Determining operating range of hyperparameters

Methods, systems, and computer program products for determining optimal compute resources for distributed batch based optimization applications are provided herein. A method includes obtaining a size of an input dataset, a size of a model, and a set of batch sizes corresponding to a job to be processed using a distributed computing system; computing, based at least in part on the set of batch sizes, one or more node counts corresponding to a number of nodes that can be used for processing said job; estimating, for each given one of the node counts, an execution time to process the job based on an average computation time for a batch of said input dataset and an average communication time for said batch of said input dataset; and selecting, based at least in part on said estimating, at least one of said node counts for processing the job.

Determining optimal compute resources for distributed batch based optimization applications

A computer generates labels for machine learning algorithms by retrieving, from a data storage circuit, multiple label sets that contain labels that each classify data points in a corpus of data. A graph is generated that includes a plurality of edges, each edge between two respective labels from different label sets of the multiple label sets. Weights are determined for the plurality of edges based upon a consistency between data points classified by two labels connected by the edges. An algorithm is applied that groups labels from the multiple label sets based upon the weights for the plurality of edges. Data points are identified from the corpus of data that represent conflicts within the grouped labels. An electronic message is transmitted in order to present the identified data points to entities for further classification. A new label set is generated using the further classification received from the entities.

Labeling of data for machine learning

Methods, systems, and computer program products for determining intermittent renewable energy penetration limits in a grid are provided herein. A computer-implemented method includes determining a load forecast for an electrical network based on historical load data and future step load information; determining the maximum size of renewable energy sources that can be added to the network based on the load forecast, parameters pertaining to the network, information pertaining to the renewable energy sources, and information pertaining to non-renewable energy sources; determining a storage component size for storing renewable energy generated by the renewable energy sources based on the maximum size of the renewable energy sources and constraints associated with the network; determining a minimum loading level of the non-renewable energy sources based on the load forecast, intermittency data associated with the renewable energy sources and the parameters pertaining to the network; and configuring the network based on the determinations.

Determining intermittent renewable energy penetration limits in a grid

Methods, systems, and computer program products for suggesting sensor placements are provided herein. A method includes simulating the concentration of a pollutant over an area covered by a sensor network grid by applying models to an emission inventory corresponding to the area; computing any discrepancy between the simulated concentration and historical data; applying scaling vectors to the first version to generate scaled versions of the emission inventory; re-simulating the concentration of the pollutant by applying the models to the scaled versions, and computing any discrepancy between the re-simulated concentrations and the historical data; identifying each scaling vector that corresponds with a discrepancy within a pre-determined range; applying the identified scaling vectors to an emission inventory value of each of the grid blocks to generate scaled emission inventory values; determining the grid blocks having high variability within scaled emission inventory values; and suggesting a sensor placement at the determined grid blocks.

Suggesting sensor placements for improving emission inventory

A method comprising using at least one hardware processor for: receiving a topic under consideration (TUC) and a set of claims referring to the TUC; identifying semantic similarity relations between claims of the set of claims; clustering the claims into a plurality of claim clusters based on the identified semantic similarity relations, wherein said claim clusters represent semantically different claims of the set of claims; and generating a list of non-redundant claims comprising said semantically different claims.

Semantic merge of arguments

Identification of an intent of a conversation can be useful for real-time or post-processing purposes. According to example embodiments, a method, and corresponding apparatus of identifying at least one intent-bearing utterance in a conversation, comprises determining at least one feature for each utterance among a subset of utterances of the conversation; classifying each utterance among the subset of utterances, using a classifier, as an intent classification or a non-intent classification based at least in part on a subset of the at least one determined feature; and selecting at least one utterance, with intent classification, as an intent-bearing utterance based at least in part on classification results by the classifier. Through identification of an intent bearing utterance, a call center for example, can provide improved service for callers through, for example, more effective directing of a call to a live agent.

System and method for identification of intent segment(s) in caller-agent conversations

Growing community of inventors

New Delhi, India

Ashish Verma