Abstract: A controller includes a feature quantity generation unit that generates, from data associated with a control target, a feature quantity appropriate for detecting an abnormality in the control target, a machine learning unit that performs machine learning using the feature quantity, an abnormality detection unit that detects the abnormality based on an abnormality detection parameter determined from a learning result of the machine learning, and the feature quantity, an instruction unit that instructs the abnormality detection unit to detect the abnormality, and a data compression unit that compresses data about the feature quantity and provides the compressed data to the machine learning unit and the abnormality detection unit. The instruction unit transmits a request for detecting the abnormality to the abnormality detection unit. The abnormality detection unit detects the abnormality without returning a response to the request.

Abstract: The disclosed computer-implemented method for file system metadata analytics may include (i) creating a set of training data to train a machine learning model to analyze tokens that describe files within a file system, the set of training data comprising a first set of vectors, wherein each vector represents tokens that describes files that are frequently accessed by a common set of users, and a second set of vectors, wherein each vector represents tokens that describes files with common file path ancestors, (ii) training, using the set of training data, the machine learning model, (iii) determining, by providing at least one input token to the machine learning model, that the input token is related to at least one additional token, and (iv) performing an action responsive to observing the input token and involving the additional token and the file system. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A technology to build emulated nervous systems is presented here, as well as the interface method for operating the emulated nervous system. The technology provides for inclusion of neuroanatomically accurate definitions organized hierarchically. This permits a highly realistic nervous system to be created and interact with its surrounding environment.

Abstract: Aspects described herein may allow for the application of stochastic gradient boosting techniques to the training of deep neural networks by disallowing gradient back propagation from examples that are correctly classified by the neural network model while still keeping correctly classified examples in the gradient averaging. Removing the gradient contribution from correctly classified examples may regularize the deep neural network and prevent the model from overfitting. Further aspects described herein may provide for scheduled boosting during the training of the deep neural network model conditioned on a mini-batch accuracy and/or a number of training iterations. The model training process may start un-boosted, using maximum likelihood objectives or another first loss function.

Abstract: Disclosed herein are system, method, and computer program product embodiments for simulating users for testing websites using a trained model. Simulating users involves training a model to generate realistic synthetic clickstreams that emulate actual clickstreams. Synthetic clickstreams may include simulated mouse actions and keystrokes and the results of the testing may be used to improve the design and functioning of the tested website.

Abstract: A learning method of a CNN (Convolutional Neural Network) for monitoring one or more blind spots of a monitoring vehicle is provided. The learning method includes steps of: a learning device, if training data corresponding to output from a detector on the monitoring vehicle is inputted, instructing a cue information extracting layer to uses class information and location information on a monitored vehicle included in the training data, thereby outputting cue information on the monitored vehicle; instructing an FC layer for monitoring the blind spots to perform neural network operations by using the cue information, thereby outputting a result of determining whether the monitored vehicle is located on one of the blind spots; and instructing a loss layer to generate loss values by referring to the result and its corresponding GT, thereby learning parameters of the FC layer for monitoring the blind spots by backpropagating the loss values.

Abstract: A method of selecting and presenting content on a first system based on user preferences learned on a second system is provided. The method includes receiving a user's input for identifying items of the second content system and, in response thereto, presenting a subset of items of the second content system and receiving the user's selection actions thereof. The method includes analyzing the selected items to learn the user's content preferences for the content of the second content system and determining a relationship between the content of the first and second content systems to determine preferences relevant to items of the first content system. The method includes, in response subsequent user input for items of the first content system, selecting and ordering a collection of items of the first content system based on the user's learned content preferences determined to be relevant to the items of the first content system.

Abstract: Embodiments relate to performing inference, such as object recognition, based on sensory inputs received from sensors and location information associated with the sensory inputs. The sensory inputs describe one or more features of the objects. The location information describes known or potential locations of the sensors generating the sensory inputs. An inference system learns representations of objects by characterizing a plurality of feature-location representations of the objects, and then performs inference by identifying or updating candidate objects consistent with feature-location representations observed from the sensory input data and location information. In one instance, the inference system learns representations of objects for each sensor. The set of candidate objects for each sensor is updated to those consistent with candidate objects for other sensors, as well as the observed feature-location representations for the sensor.

Abstract: Exemplary embodiments relate to techniques for integrating common sense into a machine learning (ML) system. In contrast to existing machine learning algorithms that search for statistical correlations between concepts, exemplary embodiments attempt to learn the semantic relationships or causality between the concepts. This may be accomplished by training an algorithm or data structure to learn similar vector representations of words present in the same context (e.g., that are present together in the same sentence). The resulting AI/ML, structure may be used to guide the generation of a causal graph having predictive capabilities. This causal graph may represent semantic relationships and/or causation between concepts, and hence may be employed to introduce a degree of common sense in the machine learning system.

Abstract: The system may be configured to perform operations including identifying, by a processor, personally identifiable information (PII) within a data model based on processing rules, to create identified PII, wherein the data model comprises entity information about an entity; comparing the identified PII with established PII in a standard data bank; validating the identified PII in response to the identified PII matching the established PII, to create validated PII; and marking the validated PII with a PII marker in response to the validating the identified PII.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for hierarchical weight-sparse convolution processing are described. An exemplary method comprises: obtaining an input tensor and a plurality of filters at a convolution layer of a neural network; segmenting the input tensor into a plurality of sub-tensors and assigning the plurality of sub-tensors to a plurality of processors; generating, for each of the plurality of filters, a hierarchical bit representation of a plurality of non-zero weights in the filter, wherein the hierarchical bit representation comprises a plurality of bits indicating whether a sub-filter has at least one non-zero weight, and a plurality of key-value pairs corresponding to the plurality of non-zero weights in the filter; identifying, based on the hierarchical bit representation, one or more of the plurality of non-zero weights and corresponding input values from the assigned sub-tensor to perform multiply-and-accumulate (MAC) operations.

Abstract: A machine learning of response selection to structured data input enables a machine to flexibly and responsively actively engage with a response recipient through a device, such as any electronic device connected to a data network. In at least one embodiment, the response selection module improves response selection to the structure data input by initially filtering a library of templates to identify candidate templates that best respond to the input. In at least one embodiment, the response selection module ranks the identified candidate templates to provide the response to the device. The response selection module learns by receiving feedback, such as a linked recipient action result signal.

Abstract: A computer-implemented method of generating a response based on a physical signal, a non-transitory memory and a system to implement the method is described. The method includes detecting a physical signal by an image sensor and a sound sensor; processing the detected physical signal by a corresponding first deep neural network; storing the processed signal as processed data in individual corresponding memory units; connecting the individual corresponding memory units with a second deep neural network to form one or more cognition units; generating, by the one more cognitions units, an expression, from a signal produced by the second deep neural network; and converting, by a third deep neural network, the expression into an output for an output device.

Abstract: Methods, apparatus, and system to determine a result of a diagnostic test strip comprising a machine learning transformer architecture which parallel processes input.

Abstract: A system and method for recommending tags for a multimedia content element to be tagged. The method includes obtaining a plurality of signatures for the multimedia content element to be tagged, wherein each of the generated signatures represents a concept, wherein each concept is a collection of signatures and metadata representing the concept; correlating between the plurality of signatures to determine at least one context of the multimedia content element to be tagged; searching for at least one contextually related multimedia content element, wherein each contextually related multimedia content element matches at least one of the determined at least one context; and identifying at least one tag, wherein each identified tag is associated with at least one of the at least one contextually related multimedia content element; generating a recommendation including the identified at least one tag.

Abstract: A computer platform implements a precision agriculture system that predicts output conditions, such as diseases, salt damage, soil problems, water leaks and generic anomalies, for orchards under analysis. The computer platform stores site and crop datasets and processed satellite image for the orchards. An orchard data learned model predicts a propensity for existence of output conditions associated with the permanent crops based on the data values for the variables of the site and crop datasets. Also, a satellite model predicts a propensity for existence of the output conditions at the orchard based on processed satellite images. A precision agriculture management model is disclosed that integrates the orchard data learned model with the satellite model to accurately predict the output conditions.

Abstract: Techniques for predicting a user response to content are described. According to various embodiments, a configuration file is accessed, where the configuration file includes a user-specification of raw data accessible via external data sources and raw data encoding rules. In some embodiments, the raw data includes raw member data associated with a particular member and raw content data associated with a particular content item. Thereafter, source modules encode the raw data from the external data sources into feature vectors, based on the raw data encoding rules. An assembler module assembles one or more of the feature vectors into an assembled feature vector, based on user-specified assembly rules included in the configuration file. A prediction module performs a prediction modeling process based on the assembled feature vector and a prediction model, to predict a likelihood of the particular member performing a particular user action on the particular content item.

Abstract: An adversarial example detection method includes: acquiring training examples and training example labels corresponding thereto, wherein the training example labels comprises normal examples and adversarial examples; inputting the training examples into a target model to obtain a first predicted score vector of the training examples; adding a random perturbation at N times to the training examples to obtain N groups of comparative training examples; respectively inputting the N groups of comparative training examples into the target model to obtain a second predicted score vector of each group of comparative training examples; constructing feature data according to the first predicted score vector and the second predicted score vector of each group of comparative training examples; training a classification model according to the feature data and the training example labels corresponding to the feature to obtain a detector; and detecting input test data according to the detector.

Abstract: Systems, apparatuses and methods may provide for technology that conducts a first timing measurement of a blockage timing of a first window of the training of the neural network. The blockage timing measures a time that processing is impeded at layers of the neural network during the first window of the training due to synchronization of one or more synchronizing parameters of the layers. Based upon the first timing measurement, the technology is to determine whether to modify a synchronization barrier policy to include a synchronization barrier to impede synchronization of one or more synchronizing parameters of one of the layers during a second window of the training. The technology is further to impede the synchronization of the one or more synchronizing parameters of the one of the layers during the second window if the synchronization barrier policy is modified to include the synchronization barrier.

Abstract: A neuromorphic weight cell (NWC) including a resistor ladder including a plurality of resistors connected in series, and a plurality of shunting nonvolatile memory (NVM) elements, each of the shunting NVM elements being coupled in parallel to a corresponding one of the resistors.