Abstract: The method provides for one or more processor receiving on a personal device, a mixture of sounds within a sound stream from multiple sources. The one or more processors identifying one or more sounds of the mixture of sounds from the multiple sources, based on a sound separation technique. The one or more processors displaying on a user interface of the personal device an icon corresponding respectively to a classification of the one or more sounds identified from the multiple sources. The one or more processors receiving a selection of a sound from the mixture of the multiple sounds, based on an action by a user of the personal device selecting the icon displayed on the user interface of the personal device, and the one or more processors recording the sound from the mixture of the multiple sounds selected by the user.

Abstract: A computer device identifies (i) a dataset, (ii) a set of output class determinations made for data entries of the dataset by a computer decision algorithm, and (iii) an undesired disparity between output class determinations resulting from a first value of a first attribute of the dataset and output class determinations resulting from a second value of the first attribute. The computing device determines a value of a second attribute of the dataset is contributing to the undesired disparity by: providing an association rule mining model (i) a first group of the data entries having the first value of the first attribute, and (ii) a second group of the data entries having the second value of the first attribute, and selecting the value of the second attribute from a set of candidate attribute values produced by the association rule mining model based, at least in part, on a lift calculation.

Abstract: Methods, systems, and computer program products for determining model-related bias associated with training data are provided herein. A computer-implemented method includes obtaining, via execution of a first model, class designations attributed to data points used to train the first model; identifying any of the data points associated with an inaccurate class designation and/or a low-confidence class designation; training a second model using the data points from the dataset, but excluding the identified data points; determining bias related to at least a portion of those data points used to train the second model by: modifying one or more of the data points used to train the second model; executing the first model using the modified data points; and identifying a change to one or more class designations attributed to the modified data points as compared to before the modifying; and outputting identifying information pertaining to the determined bias.

Abstract: A software defined data security level method, computer program product, and data processing system. One embodiment may comprise intercepting, by a processor at a data security layer, an input/output (IO) request from a local software application, wherein the IO request includes a header and a data payload, analyzing, by the processor at the data security layer, the data payload of the IO request relative to a service level agreement (SLA), assigning, by the processor at the data security layer, a security level to the IO request based on the analysis.

Abstract: A method, system, and computer program product for generating lineage events of machine learning models. The method may include identifying a machine learning model with missing lineage. The method may also include generating a creation event and deployment event for the machine learning model. The method may also include generating a version change event for the machine learning model. Generating the version change event may include identifying one or more predicted data points with a low model confidence; rescoring the one or more predicted data points based on the machine learning model at a second time period; determining that the updated one or more predicted data points are significantly different than the one or more predicted data points; and determining that there is a new version of the machine learning model. The method may also include creating a lineage path for the machine learning model.

Abstract: A processor may receive an original dataset. The processor may segment, automatically, the original dataset into a plurality of data groups. The plurality of data groups may include a model training dataset and a holdout dataset. The processor may generate a model with the model training dataset. The processor may validate the model with the holdout dataset.

Abstract: A processor may receive a record. The record may include one or more segments of text. The processor may automatically generate a first summary of the record. The processor may determine an overall bias of the first summary. The overall bias of the first summary may be identified from one or more instances of bias in the first summary. The processor may generate a second summary of the record. The second summary of the record may include an indicator of the overall bias of the first summary. The indicator may include a description of a type of overall bias of the first summary and a numerical value of the overall bias of the first summary. The processor may determine an overall bias of the second summary. The processor may display the second summary of the record to a user.

Abstract: In an approach for bypassing security vulnerable and anomalous devices in a multi-device workflow, a processor monitors behavior and network traffic of a plurality of smart devices within a multi-smart device system. A processor identifies a first smart device of the plurality of smart devices with at least one of a security vulnerability and an anomaly. A processor identifies a multi-smart device workflow that includes the first smart device. A processor identifies a function of the first smart device within the multi-smart device workflow. A processor determines whether an alternative smart device can replace the first smart device within the multi-smart device workflow. Responsive to resolution of the at least one of the security vulnerability and the anomaly, a processor re-establishes the workflow with the first smart device.

Abstract: A computer-implemented method for refining dataset to accurately represent output of an artificial intelligence model includes generating a plurality of data points used to interpret a decision of an artificial intelligence model. A subset of data points from the generated plurality of data points satisfying one or more constraints is identified. A linear model is applied on the identified subset of data points satisfying the one or more constraints. One or more insights illustrating the decision of the artificial intelligence model is generated.

Abstract: A source of bias identification (SoBI) tool is provided that identifies sources of bias in a dataset. A bias detection operation is performed on results of a computer model, based on an input dataset, to generate groupings of values for a protected attribute corresponding to a detected bias in the operation of the computer model. The SoBI tool generates a plurality of sub-groups for each grouping of values. Each sub-group comprises an individual value, or a sub-range, for the protected attribute. The SoBI tool analyzes each of the sub-groups in the plurality of sub-groups, based on at least one source of bias identification criterion, to identify one or more sources of bias in the input dataset. The SoBI tool outputs a bias notification to an authorized computing device specifying the one or more sources of bias in the input dataset.

Abstract: A method, computer system, and a computer program product for utilizing provenance data to improve machine learning is provided. Embodiments of the present invention may include collecting provenance data. Embodiments of the present invention may include identifying model quality improvements based on the collected provenance data. Embodiments of the present invention may include identifying related models based on the collected provenance data. Embodiments of the present invention may include recommending model quality improvements to a user.

Abstract: One embodiment provides a method, including: generating a plurality of ontologies wherein each ontology is generated by: monitoring interactions of a user with lineage information, wherein the monitoring comprises monitoring (i) filter interactions and (ii) access interactions; aggregating the monitored interactions of the user with monitored interactions of other users having a given business role; and generating an ontology for the given business role, wherein the subset comprises (i) event types, (ii) event constraints, (iii) event metadata, and (iv) event context; and upon a user having one of the plurality of business roles accessing lineage information on the data platform, providing a subset of the lineage information.

Abstract: Embodiments relate to a system, program product, and method for generating an enhanced feature catalog for a predictive model. The embodiments disclosed herein include capturing predictive model design time information including training data lineage metadata to determine the features of the training data, model design time measurements, and model design time metadata. Once the predictive model is built, the training data lineage metadata is used to capture the features that will be maintained within a feature catalog. The model design time measurements and model design time metadata provide further correlation between the predictive model and the features. Runtime metrics on the predictive model create additional correlations between the captured data and metadata with the features in the feature catalog to expeditiously identify the relevant features of the predictive model.

Abstract: One embodiment provides a method, including: generating a plurality of ontologies wherein each ontology is generated by: monitoring interactions of a user with lineage information, wherein the monitoring comprises monitoring (i) filter interactions and (ii) access interactions; aggregating the monitored interactions of the user with monitored interactions of other users having a given business role; and generating an ontology for the given business role, wherein the subset comprises (i) event types, (ii) event constraints, (iii) event metadata, and (iv) event context; and upon a user having one of the plurality of business roles accessing lineage information on the data platform, providing a subset of the lineage information.

Abstract: Embodiments are disclosed for a method for machine-learning model accuracy. The method includes generating prediction training data based on training predictions and corresponding probabilities of the training predictions. A classifier of a machine-learning model generates the training predictions. The method also includes training a prediction accuracy model to determine whether the training predictions generated by the machine-learning model are correct. Additionally, the method includes generating predictions in response to corresponding client transactions for the machine-learning model. Further, the method includes determining whether the predictions are accurate using the prediction accuracy model. Also, the method includes providing client predictions corresponding to the client transactions based on the determination.

Abstract: A system includes a memory having instructions therein and at least one processor in communication with the memory. The at least one processor is configured to execute the instructions to run a machine learning base model on input data to generate base model prediction data and run a machine learning error prediction model on the input data to generate error prediction data. The at least one processor is configured to execute the instructions to generate predicted correct base model prediction data based on the base model prediction data and the error prediction data. The at least one processor is configured to execute the instructions to generate confusion values data based on the base model prediction data and the predicted correct base model prediction data. The at least one processor is also configured to execute the instructions to generate base model accuracy fairness metrics data based on the confusion values data.

Abstract: According to one embodiment of the present invention, a system for modifying content associated with an item comprises at least one processor. Features of interest of the item to a plurality of different groups are determined based on user comments produced by members of the plurality of different groups. The members within each group have a common characteristic. The features of interest to each group within the content associated with the item are identified, and the content associated with the item is modified by balancing the features of interest to the plurality of different groups within the content associated with the item. Embodiments of the present invention further include a method and computer program product for modifying content associated with an item in substantially the same manner described above.

Abstract: An article is automatically augmented. The article and one or more comments are received. Comment elements are extracted from the one or more comments, and article elements are extracted from the article. Alignment scores are generated for comment-article pairs based on the extracted comment and article elements. Further, it is determined that at least one comment-article pair has an alignment score at or above a threshold alignment score. At least one augmentation feature is then generated.

Abstract: A computer-implemented method according to one embodiment includes identifying an occurrence of accuracy drift by a trained model; identifying data associated with the accuracy drift, utilizing a drift detection model (DDM) constructed for the trained model; applying the data associated with the accuracy drift to a decision tree to determine a feature space and specific subset of the data causing the accuracy drift; analyzing a distribution of features within the feature space for the specific subset of the data causing the accuracy drift to determine specific features of the data causing the accuracy drift; and returning the specific features of the data causing the accuracy drift.

Abstract: Managing notifications is provided. Personal monitoring system inputs corresponding to each member of a defined group performing a common task are contextually analyzed to identify a notification sequence for each respective member enabling task performance in a synchronized manner. Progress of each respective member while performing activities corresponding to the common task is analyzed using the personal monitoring system inputs to enable dynamic modification of the notification sequence and content to the members in accordance with the progress. Existence of any problem is identified during performance of activities corresponding to the common task to accordingly modify the notification sequence and content to target members for mitigation of an existing problem.