Abstract: Systems and methods for responsible AI compliance and governance management in AI Products are disclosed. The system receives a request to assess an enterprise product associated with a specific application. Further, the system may determine a plurality of datasets associated with the AI model of the enterprise product. Furthermore, the system generates a training dataset and a test dataset for the determined plurality of datasets associated with the AI model. The system generates a ranked list of recommended metrics for the enterprise product based on the generated training dataset and the test dataset. The system further determines a mitigation strategy for the enterprise product based on the generated ranked list of recommended metrics. Furthermore, the system creates a feedback loop for continuous training and tuning the AI model and the plurality of datasets based on the determined mitigation strategy.

Abstract: In some examples, machine learning and rule-based identification, anonymization, and de-anonymization of sensitive structured and unstructured data may include receiving input data that is to be masked, and determining, for the input data, at least one type '1 of entity extraction from a plurality of types of entity extractions to be performed on the input data. The at least one determined type of entity extraction may be performed on the input data, and at least one entity may be extracted from the input data. At least one replacement strategy may be determined from a plurality of replacement strategies for the at least one extracted entity. Further, the at least one determined replacement strategy may be applied to the at least one extracted entity to generate masked data.

Abstract: In a training phase, training data may be used to train a supervised machine learning prediction model and an unsupervised machine learning segmentation model. Then, in a testing phase, the supervised machine learning prediction model may be used to predict a target outcome for a test data observation. Also, the unsupervised machine learning segmentation model may be used to evaluate the novelty of the test data observation relative to the training data.

Abstract: Techniques and mechanisms described herein provide automated processes for integrating supervised and unsupervised classification results of a test data observation with training data observations in a feature space. Novelty of the test data observation relative to the feature space may be measured using one or more distance metrics. Novelty of a test data observation may be further refined by comparison to a confusion matrix segment determined based on a supervised model. Based on the novelty information, the supervised and/or unsupervised models may be updated, for instance via incremental or batch training.

Abstract: One or more structural equations modeling a physical process over time may be sampled using simulated parameter values to generate input data signal values. A noise generator may be applied to the input data signal values to generate noise values. The noise values and the input data signal values may be combined to determined noisy data signal values. These noisy data signal values may in turn be used in combination with one or more states to train a prediction model.

Abstract: One or more structural equations modeling a physical process over time may be sampled using simulated parameter values to generate input data signal values. A noise generator may be applied to the input data signal values to generate noise values. The noise values and the input data signal values may be combined to determined noisy data signal values. These noisy data signal values may in turn be used in combination with one or more states to train a prediction model.

Abstract: The failure modes of mechanical components may be determined based on text analysis. For example, a word embedding may be determined based on a plurality of text documents that include a plurality of maintenance records characterizing failure of mechanical components. A vector representation for a particular maintenance record may then be determined based on the word embedding. Based on the vector representation, the particular maintenance record may then be identified as belonging to a particular failure mode out of a set of possible failure modes.

Abstract: The failure modes of mechanical components may be determined based on text analysis. For example, a word embedding may be determined based on a plurality of text documents that include a plurality of maintenance records characterizing failure of mechanical components. A vector representation for a particular maintenance record may then be determined based on the word embedding. Based on the vector representation, the particular maintenance record may then be identified as belonging to a particular failure mode out of a set of possible failure modes.

Abstract: Remaining useful life may be estimated for a machine component by training a prediction model, even when limited data from actual failures is available. Feature data such as sensor readings associated with a mechanical process may be collected over time. Such readings may be paired with estimates of remaining useful life, for instance as extracted from unstructured text of maintenance records. Such data may be used to train and test the prediction model.

Abstract: A first plurality of predictor values occurring during or before a first time interval may be received. An estimated outcome value may be determined for a second time interval by applying a prediction model via a processor to the first plurality of predictor values. A designated outcome value occurring during the second time interval and a second plurality of predictor values occurring during or before the second time interval may be received. An error value may be determined based on the estimated outcome value and the designated outcome value. A drift value for a second time interval may be determined by fitting a function to the second plurality of predictor values. The prediction model may be updated when it is determined that the drift value exceeds a designated drift threshold or that the error value exceeds a designated error threshold.

Abstract: A feature data segment may be determined by applying a feature segmentation model to a test data observation. The feature segmentation model may be pre-trained via a plurality of training data observations and may divide the plurality of training data observations into a plurality of feature data segments. A predicted target value may be determined by applying to a test data observation a prediction model pre-trained via a plurality of training data observations. One or more distance metrics representing a respective distance between the test data observation and the feature data segment along one or more dimensions may be determined. The one or more distance metrics may be represented in a user interface. An updated prediction model and an updated feature segmentation model that both incorporate the test data observation and the training data observations may be determined based on user input.

Abstract: In many industrial settings, a process is repeated many times, for instance to transform physical inputs into physical outputs. To detect a situation involving such a process in which errors are likely to occur, information about the process may be collected to determine time-varying feature vectors. Then, a drift value may be determined by comparing feature vectors corresponding with different time periods. When the drift value crosses a designated drift threshold, a predicted outcome value may be determined by applying a prediction model. Sensitivity values may be determined for different features, and elements of the process may then be updated based at least in part on the sensitivity values.

Abstract: In some examples, machine learning and rule-based identification, anonymization, and de-anonymization of sensitive structured and unstructured data may include receiving input data that is to be masked, and determining, for the input data, at least one type 1 of entity extraction from a plurality of types of entity extractions to be performed on the input data. The at least one determined type of entity extraction may be performed on the input data, and at least one entity may be extracted from the input data. At least one replacement strategy may be determined from a plurality of replacement strategies for the at least one extracted entity. Further, the at least one determined replacement strategy may be applied to the at least one extracted entity to generate masked data.

Abstract: Systems and methods are disclosed for training a previously trained neural network with incremental dataset. Original train data is provided to a neural network and the neural network is trained based on the plurality of classes in the sets of training data and/or testing data. The connected representation and the weights of the neural network is the model of the neural network. The trained model is to be updated for an incremental train data. The embodiments provide a process by which the trained model is updated for the incremental train data. This process creates a ground truth for the original training data and trains on the combined set of original train data and the incremental train data. The incremental training is tested on a test data to conclude the training and to generate the incremental trained model, minimizing the knowledge learned with the original data. Thus, the results remain consistent with the original model trained by the original dataset except the incremental train data.

Abstract: The failure modes of mechanical components may be determined based on text analysis. For example, a word embedding may be determined based on a plurality of text documents that include a plurality of maintenance records characterizing failure of mechanical components. A vector representation for a particular maintenance record may then be determined based on the word embedding. Based on the vector representation, the particular maintenance record may then be identified as belonging to a particular failure mode out of a set of possible failure modes.

Abstract: Remaining useful life may be estimated for a machine component by training a prediction model, even when limited data from actual failures is available. Feature data such as sensor readings associated with a mechanical process may be collected over time. Such readings may be paired with estimates of remaining useful life, for instance as extracted from unstructured text of maintenance records. Such data may be used to train and test the prediction model.

Abstract: A method of processing data for an artificial intelligence (AI) system includes extracting features of the data to produce a lower dimensional representation of the data points; grouping the lower dimensional representation into clusters using a clustering algorithm; comparing the classes of data points within the clusters; and identifying unrepresented, under-represented, or misrepresented data.

Abstract: Systems and methods are disclosed for training a previously trained neural network with incremental dataset. Original train data is provided to a neural network and the neural network is trained based on the plurality of classes in the sets of training data and/or testing data. The connected representation and the weights of the neural network is the model of the neural network. The trained model is to be updated for an incremental train data. The embodiments provide a process by which the trained model is updated for the incremental train data. This process creates a ground truth for the original training data and trains on the combined set of original train data and the incremental train data. The incremental training is tested on a test data to conclude the training and to generate the incremental trained model, minimizing the knowledge learned with the original data. Thus, the results remain consistent with the original model trained by the original dataset except the incremental train data.

Abstract: One variation of a method for automating negotiation of goods includes: initiating a transaction negotiation package specifying a restaurant, a reservation date and time, a selection of items, and an initial monetary offer for the set of items selected by a user; retrieving a target discount rate associated with the restaurant and an aggregate discount rate negotiated across transactions previously completed with the restaurant; selecting a maximum discount rate for the transaction negotiation package based on a difference between the preset target discount rate and the aggregate discount rate; if the initial monetary offer represents an initial discount rate greater than the maximum discount rate, serving a first monetary counter offer—representing a first counter discount rate less than the maximum discount rate—to the user; and processing payment for the selection of items according to the first monetary counter offer if the first monetary counter offer is accepted.

Abstract: Systems and methods are disclosed to automatically generate a processor architecture for a custom integrated circuit (IC) described by a computer readable code. The IC has one or more timing and hardware constraints. The system extracts parameters defining the processor architecture from a static profile and a dynamic profile of the computer readable code; iteratively optimizes the processor architecture by changing one or more parameters until all timing and hardware constraints expressed as a cost function are met; and synthesizes the generated processor architecture into a computer readable description of the custom integrated circuit for semiconductor fabrication.