International Business Machines Corporation

Marek Oszajec

Lukasz Grzegorz Cmielowski

Rafal Bigaj

Wojciech Sobala

Maksymilian Erazmus

Umit M Cakmak

Pawel Slowikowski

Kamila Baron-Palucka

A first and second scoring endpoint with payload logging are deployed. At the second scoring endpoint, native data and a user-generated score for the native data are received, the native data is pre-processed into readable data for the deep-learning model, and the user-generated score and the readable data are output to the first scoring endpoint, which is associated directly with the deep-learning model. A raw payload that includes the native data is output to a payload store. At the first scoring endpoint, the readable data and the user-generated score are processed by the deep-learning model, which outputs a transformed payload and a prediction, respectively, to the payload store. The raw payload is matched with the transformed payload and the prediction to produce a comprehensive data set, which is evaluated to describe a set of transformation parameters. The deep-learning model is retrained to account for the set of transformation parameters.

Data labeling for deep-learning models

A processor may acquire a trained predictive computational model from a database. The processor may apply a trained reduced complexity model to the trained predictive computational model. The trained reduced complexity model may be associated with the trained predictive computational model. The processor may select at least one metric. The processor may determine a quality indicator related to the at least one metric by identifying the type of the at least one metric, evaluating the output of the trained predictive computational model in relation to the type of the at least one metric, and generating, based on the evaluation of the trained predictive computational model, a threshold associated with the at least one metric. The processor may determine the accuracy of the trained predictive computational model based on the quality indicator.

Automatically determining accuracy of a predictive model

A method for training a deep learning algorithm using N-dimensional data sets may be provided. Each data set comprises a plurality of N&#x2212;1-dimensional data sets. The method comprises selecting a batch size and assembling an equally sized training batch. The samples are selected to be evenly distributed within said respective N-dimensional data sets. The method comprises also starting from a predetermined offset number, wherein the number of samples is equal to the selected batch size number, and feeding said training batches of N&#x2212;1-dimensional samples into a deep learning algorithm for the training. Upon the training resulting in a learning rate that is below a predetermined level, selecting a different offset number for at least one of said N-dimensional data sets, and going back to the step of assembling. Upon the training resulting in a learning rate that is equal or higher than said predetermined level, the method stops.

Reducing problem complexity when analyzing 3-D images

Described herein is an approach for automatically determining the semantic relatedness of documents to semantic concepts. A first text mining analysis extracts a set of reference concepts from reference documents. A second text mining analysis extracts a set of test concepts from test documents that include a mixture of new concepts and reference concepts. An extended co-occurrence matrix is computed that indicates a frequency of co-occurrence (RCCF) of each new and each reference concept in the test documents with all other new and reference concepts. The extended co-occurrence matrix is used for computing a new concept relatedness score (NCRS) for the new concepts. A document similarity score (DSS) is computed for each of the test documents by aggregating, inter alia, the NCRS of each new concept with the RCCF of each reference concept. The DSS represents the semantic relatedness of the test document to the totality of the reference concepts.

Growing community of inventors

Debica, Poland

Marek Oszajec