International Business Machines Corporation

Dzung Tien Phan

Lam Nguyen

Jayant R Kalagnanam

Thanh Lam Hoang

Gabriele Picco

Vanessa Lopez Garcia

Pavankumar Murali

Nianjun Zhou

Tsuyoshi Ide

Nam H Nguyen

Ramachandran Muralidhar

Robert J Baseman

Marco Luca Sbodio

Dharmashankar Subramanian

Yufang Hou

Fateh A Tipu

Lan Cao

Long Vu

Matthew Menickelly

Ramon Fernandez Astudillo

Young-Suk Lee

Haoran Zhu

An approach to federated learning of a machine learning model may be provided. The approach may include broadcasting hyperparameters of a machine learning model to one or more client computing devices from a primary device associated with an outer loop or an inner loop. A gradient for the loss function may be calculated at the client device if previous gradients have been sufficiently large. If gradients exceeds a threshold, the client can send the mini-batch of gradients or the difference of the mini-batch of gradients back to the primary device. A search direction may be calculated based on the full gradient of the loss function for an outer loop or the mini-batch of gradient differences for an inner loop. A learning rate step may be calculated from the search direction. The hyperparameter may be updated for the inner loop based on the learning rate.

Federated learning for training machine learning models

A computer implemented method for automatically generating an optimization model for site-wide plant optimization includes mapping a process flow diagram of a plant process to a graph comprising nodes and edges, wherein the nodes represent processes and the edges represent flows between processes. A behavior is learned for each node of the graph based at least on historic data of the plant process. One or more regression functions are modeled for each node to predict an output of each of the processes, wherein the one or more regression functions are modeled based on the learned behavior for each node.

Automated generation of optimization model for system-wide plant optimization

A method, computer program product and system are provided for feature engineering and synthetic data generation. A processor retrieves a plurality of data tables, where the plurality of data tables are heterogeneous in format and content. A processor trains a variational auto-encoder (VAE) model on the plurality of data tables. A processor receives an input data table. A processor generates a synthetic data table based on the input data table and the trained VAE model.

Unsupervised learning from public tabular datasets

Software that performs the following operations: (i) receiving a set of graph predictions corresponding to an input text, where graph predictions of the set of graph predictions are generated by different respective machine learning models; (ii) blending the graph predictions of the set of graph predictions to generate a plurality of candidate blended graphs, where nodes and edges of the candidate blended graphs have respective selection metric values, generated using a selection metric function, that meet a minimum threshold; and (iii) selecting as an output blended graph a candidate blended graph of the plurality of candidate blended graphs having a highest total combination of selection metric values among the plurality of candidate blended graphs.

Blending graph predictions

Embodiments are provided for unsupervised learning of domain specific knowledge graph from textual data and language generation from knowledge graph via reinforcement learning in a computing system by a processor. Unstructured data is automatically parsed into one or more knowledge graphs based on the unstructured data and a list of candidate relations using a first machine learning model. Text data is generated from the one or more knowledge graphs using a second machine learning model.

Providing a semantic encoding and language neural network

Embodiments are provided for generating a reasonable language model learning for text data in a knowledge graph in a computing system by a processor. One or more data sources and one or more triples may be analyzed from a knowledge graph. Training data having one or more candidate labels associated with one or more of the triples may be generated. One or more reasonable language models may be trained based on the training data.

Reasonable language model learning for text generation from a knowledge graph

In some implementations, an optimization system may obtain health information identifying different measures of health of an asset. The health information identifies end of life information regarding an end of life curve of the asset and an effective age of the asset. The optimization system may determine, based on the health information, a hazard curve for the asset. The hazard curve indicates a predicted failure rate of the asset over a period of time. The optimization system may provide the hazard curve and the effective age of the asset as inputs to an optimization model. The optimization system may use the optimization model to determine a particular time for replacing the asset, wherein the particular time is determined based on the hazard curve and the effective age. The optimization system may cause the asset to be replaced at the particular time.

Asset replacement optimization based on predicted risk of failure

In example aspects of this disclosure, a method includes generating, by one or more computing devices, a parametric model that expresses condition states for each of a plurality of assets, and the probability of the assets transitioning between the condition states; generating, by the one or more computing devices, stochastic degradation predictions of a group of the assets, based on the condition states and the probability of transitioning between the condition states for at least some of the assets; and generating, by the one or more computing devices, a maintenance schedule based on: the stochastic degradation predictions of the group of the assets, costs of corrective maintenance for assets in a failed state, and costs of scheduled maintenance for the assets.

Intelligent dynamic condition-based infrastructure maintenance scheduling

A relationship between an input, a set-point of a plurality of processes and an output of a corresponding process is learned using machine learning. A regression function is derived for each process based upon historical data. An autoencoder is trained for each process based upon the historical data to form a regularizer and the regression functions and regularizers are merged together into a unified optimization problem. System level optimization is performed using the regression functions and regularizers and a set of optimal set-points of a global optimal solution for operating the processes is determined. An industrial system is operated based on the set of optimal set-points.

Prediction and operational efficiency for system-wide optimization of an industrial processing system

In an approach for compressing a neural network, a processor receives a neural network, wherein the neural network has been trained on a set of training data. A processor receives a compression ratio. A processor compresses the neural network based on the compression ratio using an optimization model to solve for sparse weights. A processor re-trains the compressed neural network with the sparse weights. A processor outputs the re-trained neural network.

Compression of deep neural networks

A computer-implemented method for a shuffling-type gradient for training a machine learning model using a stochastic gradient descent (SGD) includes the operations of uniformly randomly distributing data samples or coordinate updates of a training data, and calculating the learning rates for a no-shuffling scheme and a shuffling scheme. A combined operation of the no-shuffling scheme and the shuffling scheme of the training data is performed using a stochastic gradient descent (SGD) algorithm. The combined operation is switched to performing only the shuffling scheme from the no-shuffling scheme based on one or more predetermined criterion; and training the machine learning models with the training data based on the combined no-shuffling scheme and shuffling scheme.

Shuffling-type gradient method for training machine learning models with big data

Anomaly detection and remedial recommendation techniques for improving the quality and yield of microelectronic products are provided. In one aspect, a method for quality and yield improvement via anomaly detection includes: collecting time series sensor data during individual steps of a semiconductor manufacturing process; calculating anomaly scores for each of the individual steps using a predictive model; and implementing changes to the semiconductor manufacturing process based on the anomaly scores. A system for quality and yield improvement via anomaly detection is also provided.

Anomaly detection and remedial recommendation

A first dependency graph is constructed based on a first data set by solving an objective function constrained with a maximum number of non-zeros and formulated with a regularization term comprising a quadratic penalty to control sparsity. The quadratic penalty in constructing the second dependency graph is determined as a function of the first data set. A second dependency graph is constructed based on a second data set by solving the objective function constrained with the maximum number of non-zeros and formulated with the regularization term comprising a quadratic penalty. The quadratic penalty in constructing the second dependency graph is determined as a function of the first data set and the second data set. An anomaly score is determined for each of a plurality of sensors based on comparing the first dependency graph and the second dependency graph, nodes of which represent sensors.

Learning sparsity-constrained gaussian graphical models in anomaly detection

A computer-implemented method for controlling a production system includes mapping, by a controller, the production system as a directed acyclic graph. The production system can include multiple plants that are represented as nodes and relations between the plants represented by edges of the directed acyclic graph. The method further includes generating, by the controller, a regression model for each of the plants in the production system. The method further includes predicting, by the controller, an output of each plant based on sensor data associated from each plant. The method further includes adjusting, by the controller, one or more control variables for each plant based on a target output by using machine learning. The method further includes adjusting, by the controller, the one or more control variables for each plant to generate the target output.

Prediction optimization for system level production control

A system and method for maintaining health of a fleet of assets implementing an asset maintenance framework for collective anomaly detection that provides for a more accurate maintenance planning solution for the fleet or assets that may be prioritized. Based on a Bayesian multi-task multi-modal sparse mixture of sparse Gaussian graphical models (MTL-MM GGM), the methods combine the variational Bayes framework with (1) Laplace prior-based sparse structure learning and (2) an-based sparse mixture weight selection approach. Dual sparsity is guaranteed over both variable-variable dependency and mixture components to efficiently learn multi-modal distributions that are observed in various applications. A generated model represents the fleet-level CbM model as a combination between two model components: 1) S sets of sparse mixture weights representing individuality of the assets in the fleet; and 2) One set of sparse GGMs that are shared with the S assets to represent commonality across the S assets.

Growing community of inventors

Pleasantville, NY, United States of America

Dzung Tien Phan