Abstract: Systems, methods, and protocols for developing invasive brain computer interface (iBCI) decoders non-invasively by using emulated brain data are provided. A human operator can interact in real-time with control algorithms designed for iBCI. An operator can provide input to one or more computer models (e.g., via body gestures), and this process can generate emulated brain signals that would otherwise require invasive brain electrodes to obtain.

Abstract: A lightened neural network, method, and apparatus, and recognition method and apparatus implementing the same. A neural network includes a plurality of layers each comprising neurons and plural synapses connecting neurons included in neighboring layers. Synaptic weights with values greater than zero and less than a preset value of a variable a, which is greater than zero, may be at least partially set to zero. Synaptic weights with values greater than a preset value of a variable b, which is greater than zero, may be at least partially set to the preset value of the variable b.

Abstract: Computer network architectures for machine learning, and more specifically, computer network architectures for the automated completion of healthcare claims. Embodiments of the present invention provide computer network architectures for the automated completion of estimated final cost data for claims for healthcare clinical episodes using incomplete data for healthcare insurance claims and costs, known to date. Embodiments may use an automatic claims completion web application, with other computer network architecture components. Embodiments may include a combination of third-party databases to generate estimated final claims for pending patient clinical episodes, and to drive the forecasting models for the same, including social media data, financial data, social-economic data, medical data, search engine data, e-commerce site data, and other databases.

Abstract: A student neural network may be trained by a computer-implemented method, including: selecting a teacher neural network among a plurality of teacher neural networks, inputting an input data to the selected teacher neural network to obtain a soft label output generated by the selected teacher neural network, and training a student neural network with at least the input data and the soft label output from the selected teacher neural network.

Abstract: A micro-processor circuit and a method of performing neural network operation are provided. The micro-processor circuit is suitable for performing neural network operation. The micro-processor circuit includes a parameter generation module, a compute module and a truncation logic. The parameter generation module receives in parallel a plurality of input parameters and a plurality of weight parameters of the neural network operation. The parameter generation module generates in parallel a plurality of sub-output parameters according to the input parameters and the weight parameters. The compute module receives in parallel the sub-output parameters. The compute module sums the sub-output parameters to generate a summed parameter. The truncation logic receives the summed parameter. The truncation logic performs a truncation operation based on the summed parameter to generate a plurality of output parameters of the neural network operation.

Abstract: A computer identifies, based on sensor data from one or more sensors located in proximity to a fire site and on a corpus of firefighting knowledge, one or more firefighting goals. The computer generates, based on the one or more firefighting goals and the corpus of firefighting knowledge, one or more firefighting recommendations. The computer scores, using the corpus of firefighting knowledge, the one or more firefighting recommendations based on historical effectiveness of prior firefighting actions.

Abstract: Methods, systems, and circuits for training a neural network include applying noise to a set of training data across wordlines using a respective noise switch on each wordline. A neural network is trained using the noise-applied training data to generate a classifier that is robust against adversarial training.

Abstract: Methods are provided to classify and identify features in mass spectral data using neural network algorithms. A convolutional neural network (CNN) was trained to identify amino acids from an unknown protein sample. The CNN was trained using known peptide sequences to predict amino acid presence, diversity, and frequency, peptide length, subsequences of amino acids classified by features include aliphatic/aromatic, hydrophobic/hydrophilic, positive/negative charge, and combinations thereof. Mass spectra data of a sample unknown to the trained CNN was discretized into a one-dimensional vector and input into the CNN. The CNN models can potentially be integrated to determine the complete peptide sequence from a spectrum, thereby improving the yield of identifiable protein sequences from mass spec analysis.

Abstract: Machine learning models used in medical diagnosis should be validated after being deployed in order to reduce the number of misdiagnoses. Validation processes presented here assess a performance of the machine learning model post-deployment. In post-deployment validation, the validation process monitoring can include: (1) monitoring to ensure a model performs as well as a reference member such as another machine learning model, and (2) monitoring to detect anomalies in data. This post-deployment validation helps identify low-performing models that are already deployed, so that relevant parties can quickly take action to improve either the machine learning model or the input data.

Abstract: A method for rapid discovery of satellite behavior, applied to a pursuit-evasion system including at least one satellite and a plurality of space sensing assets. The method includes performing transfer learning and zero-shot learning to obtain a semantic layer using space data information. The space data information includes simulated space data based on a physical model. The method further includes obtaining measured space-activity data of the satellite from the space sensing assets; performing manifold learning on the measured space-activity data to obtain measured state-related parameters of the satellite; modeling the state uncertainty and the uncertainty propagation of the satellite based on the measured state-related parameters; and performing game reasoning based on a Markov game model to predict satellite behavior and management of the plurality of space sensing assets according to the semantic layer and the modeled state uncertainty and uncertainty propagation.

Abstract: Techniques are described for training and evaluating a proximal factorization machine engine. In one or more embodiments, the engine receives a set of training data that identifies a set of actions taken by a plurality of users with respect to a plurality of items. The engine generates, for a prediction model, (a) a first set of model parameters representing relationships between features of the plurality of users and the set of actions, and (b) a second set of model parameters representing interactions between different features of the plurality of users and the plurality of items. For each respective item in a plurality of items, the engine computes a probabilistic score based on the model parameters. The engine selects and presents a subset of items based on the probabilistic scores.

Abstract: The present disclosure is directed to systems and methods for the selective introduction of low-level pseudo-random noise into at least a portion of the weights used in a neural network model to increase the robustness of the neural network and provide a stochastic transformation defense against perturbation type attacks. Random number generation circuitry provides a plurality of pseudo-random values. Combiner circuitry combines the pseudo-random values with a defined number of least significant bits/digits in at least some of the weights used to provide a neural network model implemented by neural network circuitry. In some instances, selection circuitry selects pseudo-random values for combination with the network weights based on a defined pseudo-random value probability distribution.

Abstract: Embodiments for implementing mixed precision learning for neural networks by a processor. A neural network may be replicated into a plurality of replicated instances and each of the plurality of replicated instances differ in precision used for representing and determining parameters of the neural network. Data instances may be routed to one or more of the plurality of replicated instances for processing according to a data pre-processing operation.

Abstract: A 1D-CNN-based ((one-dimensional convolutional neural network)-based) distributed optical fiber sensing signal feature learning and classification method is provided, which solves a problem that an existing distributed optical fiber sensing system has poor adaptive ability to a complex and changing environment and consumes time and effort due to adoption of manually extracted distinguishable event features.

Abstract: A computational method is disclosed for the simulation of a hierarchical artificial neural network (ANN), wherein a single correlator pools, during a single time-step, two or more consecutive feed-forward inputs from previously predicted and now active neurons of one or more lower levels.

Abstract: Methods and systems for forecasting in sparse data streams. In an example embodiment, steps or operations can be implemented for mapping a time series data stream to generate forecast features using a neural network, transforming the forecast features into a space with transformed forecast features thereof using metric learning, clustering the transformed forecast features in a cluster, initializing a forecast learning algorithm with a combination of the transformed forecast features in the cluster corresponding to a sparse data stream, and displaying forecasts in a GUI dashboard with information indicative of how the forecasts were achieved, wherein the mapping, the transforming, the clustering, and the initializing together lead to increases in a speed of the forecasting and computer processing thereof.

Abstract: Techniques for auto weight scaling a bounded weight range of RPU devices with the size of the array during ANN training are provided. In one aspect, a method of ANN training includes: initializing weight values winit in the array to a random value, wherein the array represents a weight matrix W with m rows and n columns; calculating a scaling factor ? based on a size of the weight matrix W; providing digital inputs x to the array; dividing the digital inputs x by a noise and bound management factor ? to obtain adjusted digital inputs x?; performing a matrix-vector multiplication of the adjusted digital inputs x? with the array to obtain digital outputs y?; multiplying the digital outputs y? by the noise and bound management factor ?; and multiplying the digital outputs y? by the scaling factor ? to provide digital outputs y of the array.

Abstract: Techniques for statistics-aware weight quantization are presented. To facilitate reducing the bit precision of weights, for a set of weights, a quantizer management component can estimate a quantization scale value to apply to a weight as a linear or non-linear function of the mean of a square of a weight value of the weight and the mean of an absolute value of the weight value, wherein the quantization scale value is determined to have a smaller quantization error than all, or at least almost all, other quantization errors associated with other quantization scale values. A quantizer component applies the quantization scale value to symmetrically and/or uniformly quantize weights of a layer of the set of weights to generate quantized weights, the weights being quantized using rounding. The respective quantized weights can be used to facilitate training and inference of a deep learning system.

Abstract: A method may include applying, to various factors contributing to a sentiment that an end user exhibits towards an enterprise software application, a first machine learning model trained to determine, based on the factors, a sentiment index indicating the sentiment that the end user exhibits towards the enterprise software application. In response to the sentiment index exceeding a threshold value, a second machine learning model may be applied to identify remedial actions for addressing one or more of the factors contributing to the sentiment of the end user. A user interface may be generated to display, at a client device, a recommendation including the remedial actions. The remedial actions may be prioritized based on how much each corresponding factor contribute to the sentiment of the end user. Related systems and articles of manufacture are also provided.

Abstract: There are provided a neural network weight discretizing method, system and device, and a computer readable storage medium. The method includes acquiring a weight value range and a number of discrete weight states, the weight value range referring to a range of discrete weight values consisting of a maximum weight value of a current time step and a minimum weight value of the current time step, and the number of discrete weight states referring to the quantity of discrete weight states. The method also includes acquiring a weight state of a previous time step and a weight increment of the current time step and acquiring a state transfer direction by using a directional function according to the weight increment of the current time step. The method also includes acquiring a weight state of the current time step according to the weight state of the previous time step, the weight increment of the current time step, the state transfer direction, the weight value range and the number of discrete weight states.