International Business Machines Corporation

Massachusetts Institute of Technology

Wesley M Gifford

Nianjun Zhou

Pavankumar Murali

Anshul Sheopuri

Pietro Mazzoleni

Dharmashankar Subramanian

Ta-Hsin Li

Moe Z Win

Stefano Maranò

Henk A H E Wymeersch

Ashish Jagmohan

Krishna Chaitanya Ratakonda

Jing James Xu

Jayant R Kalagnanam

Bei Chen

Ying Li

Dakuo Wang

Yi-Min Chee

Anuradha Bhamidipaty

Si Er Han

Ji Hui Yang

Petros Zerfos

Rong Liu

Elham Khabiri

Patrick D Watson

Peter Daniel Kirchner

Jun Wang

Horst Cornelius Samulowitz

Nanjangud C Narendra

Dhavalkumar C Patel

Kyong Min Yeo

Nam H Nguyen

Syed Yousaf Shah

Gregory Bramble

Saket Sathe

Karthikeyan Ponnalagu

Xuan-Hong Dang

Shota Aki

Bhanukiran Vinzamuri

Long Vu

Suzanne M Rodeman

John J Ambrose

Carl Engel

Craig A Rahenkamp

A computer-implemented method, a computer program product, and a computer system for adaptive retraining of an artificial intelligence model. A computer system computes drift magnitude scores for respective drift functions. A computer system computes an aggregated data drift score for a data drift, an aggregated concept drift score for a concept drift, and an aggregated model drift score for a model drift. A computer system computes an overall drift score, based on the aggregated data drift score, the aggregated concept drift score, the aggregated model drift score, a predetermined data drift threshold, a predetermined concept drift threshold, and a predetermined model drift threshold. A computer system determines whether retraining of the artificial intelligence model is required, based on the overall drift score. A computer system performs the retraining of the artificial intelligence model, in response to determining the retraining of the artificial intelligence model is required.

Adaptive retraining of an artificial intelligence model by detecting a data drift, a concept drift, and a model drift

Embodiments of the invention are directed to collecting, by a computer system, sensor data of a manufacturing system, the sensor data being measured at intervals smaller than a time interval of a target measurement of the manufacturing system. The sensor data is determined to have a relationship to the target measurement. A synthetic target measurement is generated at an interval smaller than the time interval based on the relationship. An advance warning is automatically generated for the target measurement based on the synthetic target measurement within the interval smaller than the time interval.

Continous monitoring, advance alerts, and control of key performance indicator variables

To automate time series forecasting machine learning pipeline generation, a data allocation size of time series data may be determined based on one or more characteristics of a time series data set. The time series data may be allocated for use by candidate machine learning pipelines based on the data allocation size. Features for the time series data may be determined and cached by the candidate machine learning pipelines. Predictions of each of the candidate machine learning pipelines using at least the one or more features may be evaluated. A ranked list of machine learning pipelines may be automatically generated from the candidate machine learning pipelines for time series forecasting based upon evaluating predictions of each of the one or more candidate machine learning pipelines.

Automated time series forecasting pipeline generation

A system and a method of managing a manufacturing process includes receiving production data relating to the manufacturing process and determining an operational mode associated with the manufacturing process using historical, multivariate senor data. The method may further determine a recommended action to affect production based on the determined operational mode. The operational mode may be based on at least one of: a level of operation in a continuous flow process relating to a joint set of process variables, a representation of a joint dynamic of the set of process variables over a predefined length, and a joint configuration of an uptime/downtime of a plurality of units comprising a process flow.

Generating mode change alerts with automatic detection from sensor data

A method, computer system, and a computer program product for generating a custom corpus is provided. The present invention may include generating a domain graph. The present invention may also include gathering seed data based on the generated domain graph. The present invention may then include identifying domain related data based on the gathered seed data. The present invention may further include querying the domain related data. The present invention may also include creating word embeddings for the domain related data. The present invention may then include evaluating the domain related data.

Semi-supervised system to mine document corpus on industry specific taxonomies

A computer implemented method of administering a complex system includes receiving multivariate data from a plurality of sensors of the system in an ambient state. Event sequences in the received multivariate data are identified. The multivariate event sequences are projected to a lower stochastic latent embedding. A temporal structure of the sequences is learned in a lower latent space. A probabilistic prediction in the lower latent space is provided. The probabilistic prediction in the lower stochastic latent space is decoded to an event prediction in the ambient state.

Event time characterization and prediction in multivariate event sequence domains to support improved process reliability

A computer-implemented method for providing interpretable predictions from a machine learning model includes receiving a data structure that represents a hierarchical structure of a set of features (X) used by one or more predictive models to generate a set of predictions (Y). An interpretability model is built corresponding to the predictive models, by assigning an interpretability to each prediction Ybased on the hierarchical structure. Assigning the interpretability includes decomposing X into a plurality of partitions Xusing the hierarchical structure, wherein X=UX, N being the number of partitions. Further, each partition is decomposed into a plurality of sub-partitions using the hierarchical structure until atomic sub-partitions are obtained. A score is computed for each partition as a function of the predicted scores of the sub-partitions, wherein the predicted scores represent interactions between the sub-partitions. Further, an interpretation of a prediction is outputted.

Predictive models having decomposable hierarchical layers configured to generate interpretable results

A hybrid sensor can be generated by training a machine learning model, such as a neural network, based on a training data set. The training data set can include a first time series of upstream sensor data having forward dependence to a target variable, a second time series of downstream sensor data having backward dependence to the target variable and a time series of measured target variable data associated with the target variable. The target variable has measuring frequency which is lower than the measuring frequencies associated with the upstream sensor data and the downstream sensor data. The hybrid sensor can estimate a value of the target variable at a given time, for example, during which no actual measured target variable value is available.

Generating a hybrid sensor to compensate for intrusive sampling

The disclosure relates to extraction of rationales for studied outcome. A method comprises: grouping features as expert to align with a set of operating practices; generating interpretable features using operating rules, combining with statistical dependence analysis to bin selected features to generate favorite practice actions; grouping features as expert that combine a subset of the interpretable features to align with a set of operating practices. The method can also comprise: using a neural network or deep learning component to quantify contribution of respective experts at a consumer level applying a generic additive approach; extracting feature importance at an individual consumer-level decomposed from expert level importance; evaluating alternative, what-if, scenarios through sensitivity analysis to identify favorite practice actions; consolidating a subset of the practice actions at client or stakeholder levels; and routing respective practice actions as a function of responsibility for the set of operating practices to stakeholders or consumers.

Facilitating extraction of individual customer level rationales utilizing deep learning neural networks coupled with interpretability-oriented feature engineering and post-processing

A user trajectory graph may be constructed based on spatio-temporal data. A mobility pattern may be extracted from the user trajectory graph. Users may be clustered into groups, wherein the users in a same group possess similar feature values in the mobility pattern, and the users in different groups have different feature values, to identify personas and location sets. A distribution model may be constructed that models user timing and location preference, wherein an outcome indicates a preference for a particular time bin on a particular day for a particular location.

Automatic extraction of user mobility behaviors and interaction preferences using spatio-temporal data

Identifying one or more patterns of content consumption across multiple entities and determining an engagement action for a user of an entity based on the patterns may include receiving information associated with content, cross-industry user data associated with consumption of the content, and a given organization user data associated with consumption of the content. A first set of consumption profile vectors associated with the given organization and one or more second set of consumption profile vectors associated respectively with one or more cross-industry organizations may be generated. Information associated with a target user in the given organization may be received. A customized learning plan for the target user in the given organization may be generated based on the first set of consumption profile vectors, the one or more second set of consumption profile vectors, and the information associated with the target user.

Identifying patterns of learning content consumption across multiple entities and automatically determining a customized learning plan based on the patterns

Selecting candidates from a set of candidates by receiving a set of input parameters about a single aspect of the candidates and applying multiple scoring methods to those input parameters to compute a candidate score via each scoring method. A subset of zero or more candidates is then selected from the set of candidates by applying one or more selection criteria to the set of scores. There is not necessarily a value for every input parameter of every candidate.

Robust selection of candidates

A method of managing access to a physical mailing address using a virtual mailing address is presented. The method includes: setting up a proxy system as a server, assigning, by a server, a virtual mailing address identifier to natural or juristic person that registers with the server; linking, by the server, a physical mailing address entered by the user to the virtual mailing address identifier to generate the virtual mailing address; and enabling, by the server, a second natural person to manually access the physical mailing address when the server determines that the second user has permission to access the virtual mailing address; enabling, by the server, an institute or business (juristic person) to access the physical mailing address automatically using an Application programming interface (API) through a computer system when the server determines that the second user has permission to access the virtual mailing address.

Service to provide notification of mailing address changes

Ultra-wide bandwidth (UWB) transmission is a promising technology for indoor localization due to its fine delay resolution and obstacle-penetration capabilities. However, the presence of walls and other obstacles present a significant challenge in terms of localization, as they result in positively biased distance estimates. Measurement campaigns with FCC-compliant UWB radios can quantify effects of non-line-of-sight (NLOS) propagation. Features of waveforms measured during a campaign can be extracted for use in distinguishing between NLOS and line-of-sight situations in embodiments of the present invention. Embodiments further include classification and regression methods based on machine learning that improve the localization performance while relying solely on the received signal. Applications for systems employing an example embodiment of the invention include indoor or outdoor search and recovery with high accuracy and low cost.

Method and system for identification and mitigation of errors in non-line-of-sight distance estimation

Techniques for quantifying the quality of trend lines. A method includes defining a collection of reference data sets derived from multiple historical projects in a similar subject area as the target project, defining one or more indicators of data set applicability to a target project estimation, said defining based on multiple characteristics of the target project, evaluating applicability of each of the reference data sets based on the one or more indicators, and identifying one or more of the reference data sets as applicable for use in the target project estimation.

Quantifying the quality of trend lines

Growing community of inventors

New Canaan, CT, United States of America

Wesley M Gifford