International Business Machines Corporation

Globalfoundries Inc.

Shell USA, Inc.

Ligang Lu

Vadim Sheinin

Michael Peter Perrone

Dake He

Ashish Jagmohan

Lurng-Kuo Liu

Jack Lawrence Kouloheris

Jian Lou

Cesar Augusto Gonzales

Shan Liu

Krishna Chaitanya Ratakonda

Peter Hans Westerink

Matthias Kormaksson

Bianca Zadrozny

Alvaro Bueno Buoro

Jun Chen

Brent Paulovicks

Jorge Luis Guevara Diaz

Breno William Santos Rezende De Carvalho

Mingqi Wu

Shu Xiao

Detlef Hohl

John Tolle

Jan Limbeck

Zhou Wang

A system, method and program product for implementing a machine learning platform that processes a data map having feature and operational information. A system is disclosed that includes an interpretable machine learning model that generates a function in response to an inputted data map, wherein the data map includes feature data and operational data over a region of interest, and wherein the function relates a set of predictive variables to one or more response variables; an integration/interpolation system that generates the data map from a set of disparate data sources; and an analysis system that evaluates the function to predict outcomes at unique points in the region of interest.

Machine learning platform for processing data maps

A system, method and computer program product for seismic imaging implements a seismic imaging algorithm utilizing Reverse Time Migration (RTM) and Full Waveform Inversion (FWI) techniques requiring large communication bandwidth and low latency to convert a parallel problem into one solved using massive domain partitioning. Since in RTM and FWI, the forward wave propagation is iteratively calculated on time step increments, there are very high statistical correlation between the intermediate wavefields at the adjacent time steps. This correlation is exploited in compression of the intermediate wavefields so as to achieve much higher compression ratio and avoid the disk IO bottleneck and reduce the overall computing time substantially. Further, during simulation of a reverse wavefield, reverse predictive encoding technique is performed such that snapshots are not recovered in the reverse propagation path due to a prediction chain.

Efficient wavefield compression in seismic imaging

A system and computer program product for seismic imaging implements a seismic imaging algorithm utilizing Reverse Time Migration technique requiring large communication bandwidth and low latency to convert a parallel problem into one solved using massive domain partitioning. Since in RTM, both the forward and reverse wave propagation is iteratively calculated on time step increments, the method implements methods that includes partitioning memory between computation and intermediate results to optimize an RTM computation. The methods make maximum use of the memory to either eliminate storing the snapshot wavefield data to disk, or hide all or a significant portion of the disk I/O time. Furthermore, the schemes can provide the flexibility to vary a number of iterations (step size) for each snapshot to be kept in the memory. If any of the given conditions changes during the process, maximum usage of the available memory is ensured.

Reducing run time in seismic imaging computing

Methods and apparatus are provided for encoding video, such as surveillance video. A video frame is encoded by evaluating a level of activity in video frame relative to one or more neighboring frames; encoding the video frame using an inter-coding technique if the evaluation satisfies one or more predefined inter-coding criteria, wherein said inter-coding technique is applied to said video frame without a Group of Picture structure having a maximum number of consecutive inter-coded frames; and encoding the video frame using an intra-coding technique if the evaluation satisfies one or more predefined intra-coding criteria. The evaluating step may comprise, for example, determining whether differences between the video frame and the one or more neighboring frames exceed a predefined threshold.

Method and apparatus for encoding surveillance video

A plurality of input data streams, from a plurality of data sources, are identified. The input data streams are in a plurality of different formats. For each of the input data streams, one of a plurality of transcoders is selected, to obtain a transcoding plan. Each of the plurality of transcoders has an associated transcoding cost. A potential output for each of the plurality of transcoders has an associated storage cost, and the selecting step includes jointly reducing the overall cost, subject to appropriate constraints. The method can be employed, for example, for designing a new system or efficiently utilizing an existing system. The plurality of streams can be transcoded in accordance with the plan.

Cost and resource utilization optimization in multiple data source transcoding

A system, method and computer program product for seismic imaging implements a seismic modeling algorithm utilizing Forward Wave Inversion technique for revising Reverse Time Migration models used for sub-surface modeling. The technique requires large communication bandwidth and low latency to convert a parallel problem into one solved using massive domain partitioning. The partitioning of a velocity model into processing blocks allows each sub-problem to fit in a local cache, increasing locality and bandwidth and reducing latency. The RTM seismic data processing utilizes data that includes combined shot data, i.e., shot data selected from amongst a plurality of shots that are combined at like spatial points of the volume. An iterative approach is applied such that the correction term RTM generates at each iteration in the iterative approach is used for refining the model, and the updated model is used for generating a further refined RTM model.

Full waveform inversion using combined shot data and no scratch disk

A system, method and computer program product for generating probabilistic models of a subsurface region of the earth of interest. The system, method can be implemented efficiently to enable real-time imaging of a sub-surface structure. The system, method can provide users with the ability to assess where their subsurface images are reliable and where they are not in order to assist in the selection of low-risk, high-reward sights indicated as having oil potential for drilling. The system, method allows users to estimate a degree of uncertainty to be expected when drilling in a specific location. The knowledge of this uncertainly can be used to guide drilling in real-time to reduce the time to oil (and thereby the cost of drilling), increase the efficiency of drill maintenance and reduce the risk associated with incorrectly identifying the depth at which pressure might spike.

Probablistic subsurface modeling for improved drill control and real-time correction

Systems and techniques for directing data collection. Upon an initial data collection, the uncertainty of all or of a portion or portions of the collected data is evaluated. The collected data may be associated with a region, with portions of the collected data associated with subregions. Further data collection, including changes to or refinement of collection techniques, is undertaken based on evaluations of the uncertainty. Further data collection may be undertaken only for portions of the data for which uncertainty exceeds a threshold. Uncertainty evaluation may be performed at least in part using a model. The model may be an initial hypothesis model, and the model may be optimized as further data is collected, and the optimized model may be used to guide further data collection techniques, with iterations of data collection and model optimization being carried out concurrently.

Methods and apparatus for data collection

Systems and techniques for computational load balancing. A problem space is partitioned into subspaces and the subspaces are assigned to processing nodes. The load of nodes associated with outer subspaces is compared with the load of nodes associated with inner subspaces, and partition boundary adjustments are made based on the relative loads of outer versus inner subspaces.

Methods and apparatus for processing load balancing in distributed problem processing

A system and computer program product for seismic imaging implements a seismic imaging algorithm utilizing Reverse Time Migration (RTM) and Full Waveform Inversion (FWI) techniques requiring large communication bandwidth and low latency to convert a parallel problem into one solved using massive domain partitioning. Since in RTM and FWI, the forward wave propagation is iteratively calculated on time step increments, there are very high statistical correlation between the intermediate wavefields at the adjacent time steps. This correlation is exploited in compression of the intermediate wavefields so as to achieve much higher compression ratio and avoid the disk IO bottleneck and reduce the overall computing time substantially. Further, during simulation of a reverse wavefield, reverse predictive encoding technique is performed such that snapshots are not recovered in the reverse propagation path due to a prediction chain.

A system, method and computer program product for seismic imaging implements a seismic imaging algorithm utilizing Reverse Time Migration technique requiring large communication bandwidth and low latency to convert a parallel problem into one solved using massive domain partitioning. Several aspects of the imaging problem are addressed, including very regular and local communication patterns, balanced compute and communication requirements, scratch data handling and multiple-pass approaches. The partitioning of the velocity model into processing blocks allows each sub-problem to fit in a local cache, increasing locality and bandwidth and reducing latency. The RTM seismic data processing utilizes data that includes combined shot data, i.e., shot data selected from amongst a plurality of shots that are combined at like spatial points of the volume.

RTM seismic imaging using combined shot data

A system, method and computer program product for seismic imaging implements a seismic imaging algorithm utilizing Reverse Time Migration technique requiring large communication bandwidth and low latency to convert a parallel problem into one solved using massive domain partitioning. Several aspects of the imaging problem, including very regular and local communication patterns, balanced compute and communication requirements, scratch data handling and multiple-pass approaches. The partitioning of the velocity model into processing blocks allows each sub-problem to fit in a local cache, increasing locality and bandwidth and reducing latency.

RTM seismic imaging without scratch data storage system and method

A system and method implementing a hierarchical approach to RTM (Reverse Time Migration) seismic imaging at different granularity in space and time. An RTM seismic imaging algorithm utilizes RTM technique to convert a parallel problem into one solved using massive domain partitioning. In the method, a coarse-grain grid for the 3D volume of the geological subsurface structure under investigation is initially processed, permitting the RTM imaging process to be performed faster and produces lower level seismic image for inspection. Criteria are then applied to the first level of seismic image to determine whether to reject the image or whether a finer resolution seismic imaging is needed. In the case of finer resolution is needed, RTM resolution for the target volume is adjusted accordingly and RTM imaging process is applied with the new resolution. The process is repeated until either the image is accepted or rejected.

Growing community of inventors

New City, NY, United States of America

Ligang Lu