International Business Machines Corporation

Recruit Co., Ltd.

Wang-Chiew Tan

Peter Haas

Behzad Golshan

Chen Chen

Ignacio Gabriel Terrizzano

Ronald Mak

Lucian Popa

Paul Philip Maglio

Ronald Fagin

Yinan Li

Patricia G Selinger

Phokion Kolaitis

Yuliang Li

Xiaolan Wang

Yoshihiko Suhara

Mary Ann Roth

Cheryl Ann Kieliszewski

Aaron Traylor

Danni Ma

Jinfeng Li

Cong Yu

Ron Mak

Çağatay Demiralp

Nicole C Barberis

Wataru Hirota

Piyaphol Phoungphol

Zhengjie Miao

Çagatay Demiralp

Disclosed embodiments relate to aligning pairs of sentences. Techniques can include receiving a plurality of sentences; generating a graph for each of at least two sentences of the plurality of sentences, wherein generating a graph for each sentence of the at least two sentences comprises: identifying one or more tokens for the sentence; and connecting via edges the one or more tokens; generating a combined graph for the at least two sentences wherein generating a combined graph comprises: aligning the identified tokens of the at least two sentences of the plurality of sentences; identifying matching and non-matching tokens between the at least two sentences based on the alignment; and merging matching tokens into a combined graph node.

Systems and methods for unsupervised paraphrase mining

Disclosed embodiments relate to natural language processing. Techniques can include obtaining an encoding model, obtaining a first sentence in a first language and a label associated with the first sentence, obtaining a second sentence in a second language, encoding the first sentence and second sentence using the encoding model, determining the intent of the first encoded sentence, determining the language of the first encoded sentence and the language of the second encoded sentence, and updating the encoding model based on the determined intent of the first encoded sentence, the label, the determined language of the first encoded sentence, and the determined language of the second encoded sentence.

Systems and methods for multilingual sentence embeddings

Disclosed embodiments relate to extracting classification information from input text. Techniques can include obtaining input text, identifying a plurality of tokens in the input text, pre-training a machine learning model, determining tagging information of the plurality of tokens using a first classification layer of the machine learning model, pairing sequences of tokens using the tagging information associated with the plurality of tokens, wherein the paired sequences of tokens are determined by a second classification layer, determining one or more attribute classifiers to apply to the one or more paired sequences, wherein the attribute classifiers are determined by a third classification layer of the machine learning model, evaluating sentiments of the paired sequences, wherein the sentiments of the paired sequences are determined by a fourth classification layer of the language machine learning model, aggregating sentiments of the paired sequences associated with an attribute classifier, and storing the aggregated sentiments.

Systems and methods for semi-supervised extraction of text classification information

Disclosed embodiments relate to natural language processing. Techniques can include receiving input text, extracting, from the input text, at least one modifier and aspect pair, receiving data from a knowledgebase, based on the at least one modifier and aspect pair and commonsense data, generate one or more premise embeddings, convert the input text into tokens, generating at least one vector for one or more of the tokens based on an analysis of the tokens, combine the at least one vector with the one or more premise embeddings to create at least one combined vector, and analyze the at least one combined vector wherein the analysis generates an output indicative of a feature of the input text.

Systems and methods for enhanced review comprehension using domain-specific knowledgebases

An experiment manager is discussed for the design and execution of numerical experiments in composite simulation models, such as those created using the Smarter Planet Platform for Analysis Simulation of Health (Splash). The experiment manager independently elicits experiment-related information from each contributor of a component model, and uses this information to subsequently assist the creator of a composite model in selecting experimental factors, creating experimental designs based on these factors, and executing the experiments. This functionality permits cross-disciplinary modeling, simulation, sensitivity analysis and optimization in the setting of complex systems.

System for design and execution of numerical experiments on a composite simulation model

An aspect of synchronization of time between simulation models includes receiving a specification request for generating a set of target time-series data using a set of source time-series data, obtaining specification information relating to the set of source time-series data, obtaining specification information relating to the set of target time-series data and obtaining the source time-series data; comparing the source specification information and the target specification information to determine if the set of source time-series data are time-aligned with the set of target time-series data, and converting the set of source time-series data to the set of target time-series data upon determination that time alignment is needed.

Synchronization of time between different simulation models

A time-aware union operator is disclosed for consistent integration of time-aware data, wherein the time-aware union produces a time-aware consistent integrated view of underlying sources according to specified key constraints and policies. The implementation of time-aware union is idempotent, commutative, and associative, thus making it suitable for data integration, and it produces the same integrated outcome, modulo representation of time, regardless of the order in which sources are integrated.

System and method of integrating time-aware data from multiple sources

An aspect of handling a plurality of simulated models includes receiving from a first simulation model a stream of source data and determining data processing compatibility between the simulation model and a second-simulation model by i) obtaining metadata relating to the first-simulation models from a memory and ii) obtaining metadata relating to the second simulation model from the memory. An aspect also includes generating a transformation component using a schema mapping tool for providing compatibility between the first and second simulation models upon a determination that the compatibility does not exist between the two models. The&#x2014;transformation component is generated&#x2014;using the metadata relating to the first and second simulation models. The method also includes generating a&#x2014;stream of target data by the second simulation model via the transformation component. The second simulation model uses the stream of source data as input.

Composite simulation modeling and analysis

An aspect of combining simulation models includes a processor configured to process information relating to a plurality of simulated models. The processor collects the information to be processed related to the simulated models in at least one memory having metadata relating to the simulated models. The processor has a transformation component generated using a schema mapping tool. The transformation component detects and corrects any incompatibility between a first and a second simulation model by obtaining metadata relating to the first simulation model and the second simulation model from the memory.

Embodiments relate to a method and computer program product for generating a composite simulated model. A method includes receiving a specification request for generating a set of target time-series data from a set of source time-series data and obtaining specification information relating to the set of source time-series data, obtaining specification information relating to the set of target time-series data, and obtaining the source time-series data. The method also includes comparing the source and target specification information to determine if the set of source time-series data are time-aligned with the set of target time-series data and converting the set of source time-series data to the set of target time-series data upon determination that time alignment is needed.

A method and system for processing a schema mapping M from a source schema S to a target schema T by executing program code on a processor of a computer system. A schema mapping M' from T to S is determined via processing the schema mapping M. The schema mapping M includes at least one constraint &#x3c3;. Each constraint &#x3c3; consists of a source-to-target tuple-generating dependencies (s-t tgd). The schema mapping M&#x2032; includes at least one constraint &#x3c3;&#x2032;. Each constraint &#x3c3;&#x2032; consists of a disjunctive tgd with constants and inequalities among the constants. The schema mapping M&#x2032; is stored in at least one computer usable storage device of the computer system and/or is outputted in at least one output device of the computer system.

Growing community of inventors

San Jose, CA, United States of America

Wang-Chiew Tan