Abstract: The wastewater treatment process by using activated sludge process often appear the sludge bulking fault phenomenon. Due to production conditions of wastewater treatment process, the correlation and restriction between variables, the characteristics of nonlinear and time-varying, which lead to hard identification of sludge bulking; Sludge bulking is not easy to detect and the reasons resulting in the sludge bulking are difficult to identify, are current RBF neural network is designed for detecting and identifying the causes of sludge volume index (SVI) in this patent. The method builds soft-computing model of SVI based on recurrent RBF neural network, it has been completed to the real-time prediction of SVI concentration and better accuracy were obtained. Once the fault of sludge bulking is detected, the identifying cause variables (CVI) algorithm can find the cause variables of sludge bulking.

Abstract: A device for determining a CNN model for a database according to the present disclosure includes: a selecting unit configured to select at least two CNN models from multiple CNN models whose classification capacity is known; a fitting unit configured to fit, based on the classification capacity and first parameters of the at least two CNN models, a curve taking classification capacity and a first parameter as variables; a predicting unit configured to predict, based on the curve, a first parameter of other CNN models; and a determining unit configured to determine a CNN model applicable to the database from the multiple CNN models. With the device and the method according to the present disclosure, there is no need to train all the CNN models, thereby greatly reducing the amount of computation and simplifying the process of designing the CNN model.

Abstract: A storage controller of a machine receives training data associated with a neural network model. The neural network model includes a plurality of layers, and the machine further including at least one graphics processing unit. The storage controller trains at least one layer of the plurality of layers of the neural network model using the training data to generate processed training data. A size of the processed data is less than a size of the training data. Training of the at least one layer includes adjusting one or more weights of the at least one layer using the training data. The storage controller sends the processed training data to at least one graphics processing unit of the machine. The at least one graphics processing unit is configured to store the processed training data and train one or more remaining layers of the plurality of layers using the processed training data.

Abstract: The present disclosure provides systems and methods for on-device machine learning. In particular, the present disclosure is directed to an on-device machine learning platform and associated techniques that enable on-device prediction, training, example collection, and/or other machine learning tasks or functionality. The on-device machine learning platform can include a context provider that securely injects context features into collected training examples and/or client-provided input data used to generate predictions/inferences. Thus, the on-device machine learning platform can enable centralized training example collection, model training, and usage of machine-learned models as a service to applications or other clients.

Abstract: Embodiments are directed to a method for accelerating machine learning using a plurality of graphics processing units (GPUs), involving receiving data for a graph to generate a plurality of random samples, and distributing the random samples across a plurality of GPUs. The method may comprise determining a plurality of communities from the random samples using unsupervised learning performed by each GPU. A plurality of sample groups may be generated from the communities and may be distributed across the GPUs, wherein each GPU merges communities in each sample group by converging to an optimal degree of similarity. In addition, the method may also comprise generating from the merged communities a plurality of subgraphs, dividing each sub-graph into a plurality of overlapping clusters, distributing the plurality of overlapping clusters across the plurality of GPUs, and scoring each cluster in the plurality of overlapping clusters to train an AI model.

Abstract: An apparatus and method are provided for review-based machine learning. Included are a non-transitory memory storing instructions and one or more processors in communication with the non-transitory memory. The one or more processors execute the instructions to receive first data, generate a plurality of first features based on the first data, and identify a first set of labels for the first data. A first model is trained using the first features and the first set of labels. The first model is reviewed to generate a second model, by receiving a second set of labels for the first data, and reusing the first features with the second set of labels in connection with training the second model.

Abstract: Processors and methods for neural network processing are provided. A method in a processor including a pipeline having a matrix vector unit (MVU), a first multifunction unit connected to receive an input from the MVU, a second multifunction unit connected to receive an output from the first multifunction unit, and a third multifunction unit connected to receive an output from the second multifunction unit is provided. The method includes decoding instructions including a first type of instruction for processing by only the MVU and a second type of instruction for processing by only one of the multifunction units. The method includes mapping a first instruction for processing by the matrix vector unit or to any one of the first multifunction unit, the second multifunction unit, or the third multifunction unit depending on whether the first instruction is the first type of instruction or the second type of instruction.

Abstract: Techniques for classifying cardiac events in electrocardiogram (ECG) data. A feature set is generated by analyzing ECG data for a patient using a first phase in a machine learning architecture. A first cardiac event in the ECG data is classified based on the feature set, using the first phase in the machine learning architecture. A second cardiac event in the ECG data is classified based on the classified first cardiac event and the feature set, using a second phase in the machine learning architecture. The second cardiac event overlaps at least partially in time with the first cardiac event. Further, a plurality of feature sets, corresponding to a plurality channels of ECG data, are generated using paths in a machine learning architecture. A cardiac event in the ECG data is classified using the machine learning architecture based on the plurality of feature sets.

Abstract: A system and method for the preparation of food is provided. The system utilizes a nutritional information module which allows nutritional information to be aggregated for an entire menu along with portion information and the ability to adjust serving weight by preferred caloric value for any individual recipe. The system also utilizes a scheduler module which compiles task information for any individual recipe items per paragraph. The schedule module may have any plurality of different task times including a passive task time and/or an active task time and may time stamp the time it takes to prepare any specific recipe. The schedule may compile cook information and store information in a memory bank for analysis of the individual cook's cooking style and cook time. Moreover, the system may also provide a feedback module which uses personal time co-efficients to predict how long it should take for any particular menu choice preparation.

Abstract: The disclosed technology relates identifying causes of an observed outcome. A system is configured to receive an indication of a user experience problem, wherein the user experience problem is associated with observed operations data including an observed outcome. The system generates, based on the observed operations data, a predicted outcome according to a model, determines that the observed outcome is within range of the predicted outcome, and identifies a set of candidate causes of the user experience problem when the observed outcome is within range of the predicted outcome.

Abstract: A method for generating a target network by performing neural architecture search using optimized search space is provided. The method includes steps of: a computing device (a) if a target data is inputted into the target network, allowing the target network to apply neural network operation to the target data, to generate an estimated search vector; and (b) allowing a loss layer to calculate architecture parameter losses by referring to the estimated search vector and a ground truth search vector, and to perform backpropagation by referring to the architecture parameter losses to update architecture parameter vectors for determining final layer operations among candidate layer operations included in an optimized layer type set corresponding to the optimized search space and wherein the final layer operations are to be performed by neural blocks, within cells of the target network, arranged according to an optimized cell template corresponding to the optimized search space.

Abstract: Circuitry of a pulse generation program compiler is operable to parse pulse generation program source code comprising a declaration of a non-stream variable, a declaration of a stream variable, and one or more stream processing statements that reference the stream variable. The circuitry of the pulse generation program compiler is operable to generate, based on the declaration of the non-stream variable, a machine for execution by a quantum controller and a quantum orchestration server.

Abstract: The present invention relates generally to identifying relationships between items. Certain embodiments of the present invention are configurable to identify the probability that a certain event will occur by identifying relationships between items. Certain embodiments of the present invention provide an improved supervised machine learning system.

Abstract: Aspects of the present disclosure relate to identifying a product in a document. A computing machine accesses a product mention in a scientific or research-related text, the product mention including one or more attribute values for a plurality of attributes, each attribute being associated with either a single attribute value or no attribute value. The computing machine determines that the attribute values of the product mention correspond to two or more candidate product matches in a product directory. The computing machine identifies, based at least in part on stored data related to the scientific or research-related text, a product match from among the candidate product matches, the product match corresponding to the product mention in the scientific or research-related text. The computing machine provides an output of the product match for storage in conjunction with the product mention in the scientific or research-related text.

Abstract: An input data set with a plurality of item descriptors comprising respective time series observations is identified. A routing directive indicating a predicate to be evaluated to determine whether a particular item descriptor is to be included in a training data set for a first learning algorithm is obtained. A plurality of learning algorithms are trained using training data sets derived from the input data set according to respective routing directives, and the trained algorithms are stored.

Abstract: Data is input to one of a plurality of neural networks. Each of the plurality of neural networks is to be of a different size. A propagation time is determined for the inputted data. The propagation time relates to a time for the inputted data to propagate through one of the plurality of neural networks. One of the plurality of neural networks is selected based on the propagation time.

Abstract: A response generation apparatus ensures accurate output. A computer stores graph knowledge including a response generation module generating a response to an input document including a plurality of sentences, the graph knowledge database includes graph data that manages a structure of each type of graph knowledge, and the response generation module generates a first graph knowledge from each of the sentences; searches a second graph knowledge similar to each of the plurality of first graph knowledge while referring to the graph data on the basis of the plurality of first graph knowledge; identifies the plurality of second graph knowledge included in a dense location where a density of the second graph knowledge is high in a graph space; searches third graph knowledge for generating the response while referring to the graph data on the basis of the identified second graph knowledge; and generates the response using the third graph knowledge.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for engineering a data analytics platform using machine learning are disclosed. In one aspect, a method includes the actions of receiving data indicating characteristics of data for analysis, analysis techniques to apply to the data, and requirements of users accessing the analyzed data. The actions further include accessing provider information that indicates computing capabilities of a respective data analysis provider, analysis techniques provided by the respective data analysis provider, and real-time data analysis loads of the respective data analysis provider. The actions further include applying the characteristics of the data, the analysis techniques, the requirements of the users, and the provider information, the analysis techniques, and the real-time data analysis loads to a model.

Abstract: A method for retraining an artificial neural network trained on data from an old task includes training the artificial neural network on data from a new task different than the old task, calculating a drift, utilizing Sliced Wasserstein Distance, in activation distributions of a series of hidden layer nodes during the training of the artificial neural network with the new task, calculating a number of additional nodes to add to at least one hidden layer based on the drift in the activation distributions, resetting connection weights between input layer nodes, hidden layer nodes, and output layer nodes to values before the training of the artificial neural network on the data from the new task, adding the additional nodes to the at least one hidden layer, and training the artificial neural network on data from the new task.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for balanced-weight sparse convolution processing.