Abstract: The technology evaluates the compliance of an AI application with predefined vector constraints. The technology employs multiple specialized models trained to identify specific types of non-compliance with the vector constraints within AI-generated responses. One or more models evaluate the existence of certain patterns within responses generated by an AI model by analyzing the representation of the attributes within the responses. Additionally, one or more models can identify vector representations of alphanumeric characters in the AI model's response by assessing the alphanumeric character's proximate locations, frequency, and/or associations with other alphanumeric characters. Moreover, one or more models can determine indicators of vector alignment between the vector representations of the AI model's response and the vector representations of the predetermined characters by measuring differences in the direction or magnitude of the vector representations.

Abstract: Systems and methods for transforming data between multiple styles are provided. In one embodiment, a system is provided that includes a generator model, a discriminator model, and a preserver model. The generator model may be configured to receive data in a first style and generate converted data in a second style. The discriminator model may be configured to receive the converted data from the generator model, compare the converted data to original data in the second style, and compute a resemblance measure based on the comparison. The preserver model may be configured to receive the converted data from the generator model and compute an information measure of the converted data. The generator model may also be trained to optimize the resemblance measure and the information measure.

Abstract: The disclosed embodiments relate to a system that generates and executes a deep neural network (DNN) based on target runtime parameters. During operation, the system receives a trained original model and a set of target runtime parameters for the DNN, wherein the target runtime parameters are associated with one or more of the following for the DNN: desired operating conditions, desired resource utilization, and desired accuracy of results. Next, the system generates a context-specific model based on the original model and the set of target runtime parameters. The system also generates an operational plan for executing both the original model and the context-specific model to meet requirements of the target runtime parameters. Finally, the system controls execution of the original model and the context-specific model based on the operational plan.

Abstract: Embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and/or the like for generating an inferred document representation for a multi-section document using a machine learning model.

Abstract: The present disclosure discloses a method for deriving fault diagnosis rules of a blast furnace based on a deep neural network, which relates to the field of industrial process monitoring, modeling and simulation. Firstly, a deep neural network is used to model historical fault data of the blast furnace. Then, for each kind of fault, the process starts from the output layer of the network, wherein sub-models of nodes in the adjacent layers in the deep neural network are established by using the decision tree in sequence, and the if-then rule is derived. Finally, the if-then rules are merged layer by layer, so as to finally obtain fault diagnosis rules of the blast furnace with blast furnace process variables being the rule antecedents and with fault categories being the rule consequents.

Abstract: The present disclosure describes system and methods for accessing data from a gas oil separation plant (GOSP) facility, wherein the data includes measurements at various locations inside the GOSP facility and measurements of water cut of the GOSP facility; selecting, based on feature engineering, a subset of features corresponding to the measurements at various locations inside the GOSP facility, wherein the subset of features are more likely to impact the water cut of the GOSP facility than unselected features; and based on the subset of features, training a predictive model capable of predicting the water cut of the GOSP facility based on the measurements of water cut of the GOSP facility, wherein the training is based on, at least in part, (i) a subset of the measurements at various locations inside the GOSP facility and (ii) a subset of the measurements of water cut of the GOSP facility.

Abstract: An approach of accelerating inferences based on decision trees based on accessing one or more decision trees, wherein each decision tree of the decision trees accessed comprises decision tree nodes, including nodes grouped into one or more supersets of nodes designed for joint execution. For each decision tree of the decision trees accessed, the nodes are executed to obtain an outcome for the one or more decision trees, respectively. For each superset of the one or more supersets of said each decision tree, the nodes of each superset are jointly executed by: loading attributes of the nodes of each superset in a respective cache line of the cache memory processing said attributes from the respective cache line until an inference result is returned based on the one or more outcomes.

Abstract: This disclosure generally provides solutions for improving the performance of a custom-built, packet-switched, TPU accelerator-side communication network. Specifically a set of solutions to improve the flow-control behavior by tuning the packet buffer queues in the on-chip router in the distributed training supercomputer network are described.

Abstract: An approach to identifying architectures of machine learning models meeting a user defined constraint. The approach can receive input associated with evaluating machine learning models from a user. The approach can determine acceptable architectural templates to evaluate the machine learning models based on the input and determine a list of architectures and metrics based on a calculation of maximum neural network sizes of the acceptable architectural templates not exceeding the constraint. The approach can send the list of architectures and metrics to the user for selection.

Abstract: Systems, computer program products, and methods are described herein for automated regression testing. The present invention is configured to generate, using a regression testing engine, a second state instance map for a second version of an application; generate, using the regression testing engine, a first state instance map for a first version of the application; initiate a differential detection engine on the first state instance map and the second state instance map; determine, using the differential detection engine, one or more differential features in the second version of the application; initiate a machine learning model on the one or more differential features in the second version of the application; and classify, using the machine learning model, the one or more differential features into one or more classes.

Abstract: The disclosure herein provides methods, systems, and devices for measuring similarity of and generating recommendations for unique items. A recommendation system for generating recommendations of alternative unique items comprises an items information database, a penalty computation engine, a recommendation compilation engine, and one or more computers, wherein the penalty computation engine comprises a customizations filter, a condition filter, and a dissimilarity penalty calculator.

Abstract: Data-dependent node-to-node knowledge sharing to increase the interpretability of the activation pattern of one or more nodes in a neural network, is implemented by a set of knowledge sharing links. Each link may comprise a knowledge providing node or other source P and a knowledge receiving node R. A knowledge sharing link can impose a node-specific regularization on the knowledge receiving node R to help guide the knowledge receiving node R to have an activation pattern that is more easily interpreted. The specification and training of the knowledge sharing links may be controlled by a cooperative human-AI learning supervisor system in which a human and an artificial intelligence system work cooperatively to improve the interpretability and performance of the client system.

Abstract: There is a need for more effective and efficient predictive data analysis based at least in part on categorical input data. This need can be addressed by, for example, solutions for performing predictive data analysis that utilize at least one of categorical level merging, mutual-information-based feature filtering, feature-correlation-based feature filtering to generate training data feature value arrangements, as well as training and using categorical input machine learning models trained using the training data feature value arrangements.

Abstract: Embodiments relate to an inference method and device using a spiking neural network including parameters determined using an analog-valued neural network (ANN). The spiking neural network used in the inference method and device includes an artificial neuron that may have a negative membrane potential or have a pre-charged membrane potential. Additionally, an inference operation by the inference method and device is performed after a predetermined time from an operating time point of the spiking neural network.

Abstract: A method for operating an artificial neural network (ANN) includes quantifying a reward for executing ANN tasks in a system having multiple processing cores. A set of processing cores of the multiple processing cores is allocated to execute each of the tasks based on the reward. The ANN tasks are executed concurrently according to the processing core allocation to operate the ANN.

Abstract: Systems, methods, and computer readable media related to training and/or utilizing a neural network model to determine, based on a sequence of sources that each have an electronic interaction with a given electronic resource, one or more subsequent source(s) for interaction with the given electronic resource. For example, source representations of those sources can be sequentially applied (in an order that conforms to the sequence) as input to a trained recurrent neural network model, and output generated over the trained recurrent neural network model based on the applied input. The generated output can indicate, for each of a plurality of additional sources, a probability that the additional source will subsequently (e.g., next) interact with the given electronic resource. Such probabilities indicated by the output can be utilized in performance of further electronic action(s) related to the given electronic resource.

Abstract: A method, system, and computer program product for learning parameters of Bayesian network using uncertain evidence, the method comprising: receiving input comprising graph representation and at least one sample of a Bayesian network, the graph comprising plurality of nodes representing random variables and plurality of directed edges representing conditional dependencies, wherein each of the at least one sample comprising for each node a value selected from the group consisting of: a known value; an unknown value; and an uncertain value; and applying on the input a Bayesian network learning process configured for calculating estimates of conditional probability tables of the Bayesian network using probabilities inferred by applying on the input a Bayesian network uncertain inference process configured for performing inference in a Bayesian network from uncertain evidence.

Abstract: Some embodiments of the invention provide a method for configuring a network with multiple nodes. Each node generates an output value based on received input values and a set of weights that are previously trained to each have an initial value. For each weight, the method calculates a factor that represents a loss of accuracy to the network due to changing the weight from its initial value to a different value in a set of allowed values for the weight. Based on the factors, the method identifies a subset of the weights that have factors with values below a threshold. The method changes the values of each weight from its initial value to one of the values in its set of allowed values. The values of the identified subset are all changed to zero. The method trains the weights beginning with the changed values for each weight.

Abstract: The present invention provides a system and method for managing operations in an enterprise application. The invention includes predicting a dataset characteristic such as an incident or outage in the enterprise application based on analysis of the dataset received from distinct data sources. The invention includes data cleansing, feature extraction based on probability data analysis and classification of dataset based on data models trained on historical dataset characteristic data. The invention includes linkedchain based architecture with configurable components structuring the enterprise application.

Abstract: Techniques are disclosed for automatically retraining a machine learning model based on the performance of this model falling below a performance threshold. In some embodiments, a computer system compares output of a new machine learning model for a new set of examples with known labels for examples in the new set of examples, wherein the new set of examples includes one or more new features. In some embodiments, the computer system determines, based on the comparing, whether a current performance of the new machine learning model satisfies a performance threshold for machine learning models, where the performance threshold is based on output of a benchmark machine learning model. In some embodiments, the computer system automatically triggers, in response to determining that the current performance of the new model does not satisfy the performance threshold, retraining of the new model.