International Business Machines Corporation

Universiti Brunei Darussalam

Yogish Sabharwal

Vaibhav Saxena

Thomas George

Ashish Verma

Monu Kedia

Anamitra Roy Choudhury

Venkatesan T Chakaravarthy

Saurav Basu

Rashmi Mittal

Shakti Kapoor

John A Gunnels

Ravi Kothari

Ashok Pon Kumar Sree Prakash

Sameer Kumar

Rahul Garg

James C Sexton

Nikhil Jain

Jayaram Kallapalayam Radhakrishnan

Saurabh Goyal

Saurabh Manish Raje

Chandrasekar Radhakrishnan

Anshul Mittal

Padmanabha Venkatagiri Seshadri

Amith Singhee

Vinayaka D Pandit

Raghavendra Singh

Anthony Paul Praino

Jayanta Basak

Devasenapathi Periagraharam Seetharamakrishnan

Lloyd Alan Treinish

Subhrajit Bhattacharya

Koyel Mukherjee

Parikshit Ram

Aswin Kannan

Saritha Vinod

Shrihari Vasudevan

Saiful A Husain

James Peter Cipriani

Prashant Agrawal

Sambuddha Roy

Alind Khare

Jayasuriya M R Sarath Bandara

Vanessa Teo

Methods, systems, and computer program products for cold-start service placement over on-demand resources are provided herein. A computer-implemented method includes obtaining a performance requirement profile comprising performance requirements of a service that vary over time; determining a plurality of incarnations for the service, wherein each incarnation is associated with a level of performance provided by the incarnation for the service, resource requirements of the incarnation, and a type of computing node the incarnation is configured to execute on; identifying computing nodes having different types and different resource capacities; jointly scheduling (i) the computing nodes and (ii) one or more of the incarnations on the computing nodes over a time interval such that a cumulative level of performance of the incarnations scheduled at each timepoint in the time interval satisfies the performance requirement profile of the service.

Cold-start service placement over on-demand resources

A system, computer program product, and method are presented for performing multi-objective automated machine learning, and, more specifically, to identifying a plurality of machine learning pipelines as Pareto-optimal solutions to optimize a plurality of objectives. The method includes receiving input data directed toward one or more subjects of interest and determining a plurality of objectives to be optimized. The method also includes ingesting at least a portion of the input data through one or more machine learning (ML) models. The method further includes aggregating the plurality of objectives into one or more aggregated single objectives. The method also includes determining a plurality of Pareto-optimal solutions, thereby defining a plurality of ML pipelines that optimize the one or more aggregated single objectives. The method further includes selecting one ML pipeline from the plurality of ML pipelines.

Multi-objective automated machine learning

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

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

One embodiment provides a computer implemented method, including: receiving an application including a plurality of services to be hosted on a remote services environment; obtaining, a resource profile identifying a usage of a resource by a given service of the plurality of services over a period of time; splitting, based upon the resource profile corresponding to a given service, the given service into a plurality of service slices; selecting, for each of the plurality of service slices an incarnation fulfilling a resource demand requirement and a service-performance-offering, wherein an incarnation has a total demand value based upon a resource capacity of a node at the remote services environment; and assigning, for each of the plurality of service slices the incarnation selected to at least one node within the remote services environment based upon the resource capacity of the at least one node.

Incarnation selection for application hosting on remote services environment

Methods and systems for managing energy sources and energy consumers in an integrated system are provided. Certain subject matter presented herein relates to automatically scheduling jobs and their sub-tasks to maximize profit by comparing power source configurations and determining the best job schedule for the power source configurations. This system broadly involves two sub-problems: determining the best power source configuration and determining the best job schedule for the given power source configuration.

Management of power sources and jobs in an integrated power system

Methods, systems, and computer program products for scheduling crop transplantations are provided herein. A method includes determining one or more lifecycle constraints associated with a given crop via analysis of crop data; determining one or more cultivation conditions constraints associated with cultivating the given crop via analysis of cultivation conditions data; determining one or more weather forecast constraints associated with a given geographic area via analysis of weather forecast data; and generating a transplantation schedule for the given crop in the given geographic area based on determining a fit across (i) the one or more lifecycle constraints, (ii) the one or more cultivation conditions constraints, and (iii) the one or more weather forecast constraints.

Scheduling crop transplantations

One embodiment provides a method for determining at least one recommended vehicular travel route, the method including collecting a plurality of travel route inputs from a plurality of sources, relative to a travel route comprising a starting point and an ending point; determining at least one travel route recommendation by (i) assessing the collected plurality of travel route inputs and (ii) taking in to consideration vehicle emission impact parameters; and tracking vehicle adherence. Other variants and embodiments are broadly contemplated herein.

Travel planning based on minimizing impact of vehicular emission

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

Methods and arrangements for automating weather model configurations. Characteristics of a target geographical area are received, as are a plurality of existing weather model configurations which are implemented for undertaking weather modeling in other geographical areas. A subset of the existing weather model configurations is selected, the subset comprising configurations corresponding to geographical areas having characteristics similar to the characteristics of the target area. A weather model is run with respect to each configuration in the subset of existing weather model configurations. Based on the running of a weather model with respect to each configuration, an output set of weather model configurations is selected for undertaking weather modeling in the target area.

Automating weather model configurations

Methods and arrangements for performing synchronized collective operations. Communication calls are accepted from at least two distinct processor groups. Edge disjoint spanning paths are created over a collective comprising the processor groups, and the spanning paths are assigned to the processor groups to facilitate communication within each processor group.

Performing synchronized collective operations over multiple process groups

Growing community of inventors

New Delhi, India

Yogish Sabharwal