Abstract: Auto-generation of dialog for a skill executed by a digital assistant is performed. Details descriptive of a digital assistant persona are received from a personality studio user interface. A personality type is generated based on the details. A standard vocabulary is auto-generated using the personality type. The standard vocabulary is exported for use in a skill. Prompts are auto-generated for the skill based on the standard vocabulary.

Abstract: Systems, methods, and computer-readable media are described for determining the orientation of a target object in an image and iteratively reorienting the target object until an orientation of the target object is within an acceptable threshold of a target orientation. Also described herein are systems, methods, and computer-readable media for verifying that an image contains a target object.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for searching an autonomous vehicle sensor data repository. One of the methods includes maintaining a collection of sensor samples and, for each sensor sample, an embedding of the sensor sample; receiving a request specifying a query sensor sample, wherein the query sensor sample characterizes a query environment region; and identifying, from the collection of sensor samples, a plurality of relevant sensor samples that characterize similar environment regions to the query environment region, comprising: processing the query sensor sample through the embedding neural network to generate a query embedding; and identifying, from sensor samples in a subset of the sensor samples in the collection, a plurality of sensor samples that have embeddings that are closest to the query embedding.

Abstract: Interactions between a training server and a plurality of environment controllers are used for updating the weights of a predictive model used by a neural network executed by the plurality of environment controllers. Each environment controller executes the neural network using a current version of the predictive model to generate outputs based on inputs, modifies the outputs, and generates metrics representative of the effectiveness of the modified outputs for controlling the environment. The training server collects the inputs, the corresponding modified outputs, and the corresponding metrics from the plurality of environment controllers. The collected inputs, modified outputs and metrics are used by the training server for updating the weights of the current predictive model through reinforcement learning. A new predictive model comprising the updated weights is transmitted to the environment controllers to be used in place of the current predictive model.

Abstract: Provided is a spike neural network circuit including a synapse configured to generate an operation signal based on an input spike signal and a weight, and a neuron configured to generate an output spike signal using a comparator configured to compare a voltage of a membrane signal generated based on the operation signal with a voltage of a threshold signal, wherein the comparator includes a bias circuit configured to conditionally supply a bias current of the comparator depending on the membrane signal.

Abstract: A neural processing unit (NPU) is described. The NPU includes an NPU direct memory access (NDMA) core. The NDMA core includes a read engine having a read buffer. The NDMA core also includes a write engine having a write buffer. The NPU also includes a controller. The controller is configured to direct the NDMA core to perform hardware pre-processing of NDMA data in the read buffer and post-processing of NDMA data in the write buffer on blocks of a data stripe to process tensors in artificial neural networks.

Abstract: A data format converter rearranges data of an input image for input to a systolic array of multiply and accumulate processing elements. The image has a pixel height and a pixel width in a number of channels equal to a number of colors per pixel. The data format converter rearranges the data to a second, greater number of channels and inputs the second number of channels to one side of the systolic array. The second number of channels is less than or equal to the number of MAC PEs on the one side of the systolic array, and results in greater MAC PE utilization in the systolic array.

Abstract: A neural processing unit of a binarized neural network (BNN) as a hardware accelerator is provided, for the purpose of reducing hardware resource demand and electricity consumption while maintaining acceptable output precision. The neural processing unit may include: a first block configured to perform convolution by using a binarized feature map with a binarized weight; and a second block configured to perform batch-normalization on an output of the first block. A register having a particular size may be disposed between the first block and the second block. Each of the first block and the second block may include one or more processing engines. The one or more processing engines may be connected in a form of pipeline.

Abstract: Disclosed is a low-power and compact neuron circuit implementing a ReLU activation function including a first-layer synaptic array, a neuron transistor, a resistor, and a second-layer synaptic array. The neuron transistor is a MOS transistor having a threshold voltage-adjustable property, a gate electrode of the neuron transistor is connected to each voltage output end of the first-layer synaptic array, and a drain electrode of the neuron transistor is connected to each voltage input end of the second-layer synaptic array. Thus, it is possible to satisfy the decision computation and output of different synaptic array output values by adjusting the magnitude of the threshold voltage of the transistor. The neuron circuit requires only one transistor in cooperative connection with the first-layer synaptic array and the second-layer synaptic array to implement the ReLU activation function; therefore, a significant improvement is achieved in terms of energy efficiency, delay reduction, and space utilization.

Abstract: A neuron circuit, comprising first and second NDR devices biased each with opposite polarities, said first and second NDR devices being coupled to first and second grounded capacitors.

Abstract: Disclosed by the disclosure is a convolutional neural network on-chip learning system based on non-volatile memory, comprising: an input module, a convolutional neural network module, an output module and a weight update module. The on-chip learning of the convolutional neural network module implements the synaptic function by using the characteristic of the memristor, and the convolutional kernel value or synaptic weight value is stored in a memristor unit; the input module converts the input signal into the voltage signal; the convolutional neural network module converts the input voltage signal layer-by-layer, and transmits the result to the output module to obtain the output of the network; and the weight update module adjusts the conductance value of the memristor in the convolutional neural network module according to the result of the output module to update the network convolutional kernel value or synaptic weight value.

Abstract: A machine learning environment utilizing training data generated by customer environments. A reinforced learning machine learning environment receives and processes training data generated by independently hosted, or decoupled, customer environments. The reinforced learning machine learning environment corresponds to machine learning clusters that receive and process training data sets provided by the decoupled customer environments. The customer environments include an agent process that collects training data and forwards the training data to the machine learning clusters without exposing the customer environment. The machine learning clusters can be configured in a manner to automatically process the training data without requiring additional user inputs or controls to configured the application of the reinforced learning machine learning processes.

Abstract: We disclose computational models that alleviate the effects of human ascertainment biases in curated pathogenic non-coding variant databases by generating pathogenicity scores for variants occurring in the promoter regions (referred to herein as promoter single nucleotide variants (pSNVs)). We train deep learning networks (referred to herein as pathogenicity classifiers) using a semi-supervised approach to discriminate between a set of labeled benign variants and an unlabeled set of variants that were matched to remove biases.

Abstract: Various embodiments provide for quantizing a trained neural network with removal of normalization with respect to at least one layer of the quantized neural network, such as a quantized multiple fan-in layer (e.g., element-wise add or sum layer).

Abstract: Data may be abstracted and/or masked prior to being provided to a machine learning model for training. A machine learning model may provide a confidence level associated with a result. If the confidence level is too high, the machine learning model or an application including the machine learning model may refrain from providing the result as an output. In some examples, the machine learning model may provide a “second best” result that has an acceptable confidence level. In other examples, an error signal may be provided as the output. In accordance with examples of the present disclosure, data may be abstracted and/or masked prior to being provided to a machine learning model for training and confidence levels of results of the trained machine learning model may be used to determine when a result should be withheld.

Abstract: Systems, apparatuses and methods may provide for technology that identifies a cognitive space that is to be a compressed representation of activations of a neural network, maps a plurality of activations of the neural network to a cognitive initial point and a cognitive destination point in the cognitive space and generates a first cognitive trajectory through the cognitive space, wherein the first cognitive trajectory traverses the cognitive space from the cognitive initial point to the cognitive destination point.

Abstract: Example apparatus disclosed herein are to process a first image of a first video segment from the image capture sensor with a machine learning algorithm to determine a first score for the first image, the machine learning algorithm to detect actions associated with images, the actions associated with labels. Disclosed example apparatus are also to determine a second score for the first video segment based on respective first scores for corresponding images in the first video segment. Disclosed example apparatus are further to determine, based on the second score, whether to retain the first video segment in the memory.

Abstract: A system that includes artificial intelligence (AI) configured to identify text and images within an industrial reference. Example industrial references include electrical drawings and P&IDs. The system includes a method for training artificial intelligence model to recognize text characters and strings in addition to industrial images using a limited sample set. The use of a limited sample set improves computer performance by relying on a smaller dataset to train the model.

Abstract: A system and method implement deep learning on a mobile device to provide a convolutional neural network (CNN) for real time processing of video, for example, to color hair. Images are processed using the CNN to define a respective hair matte of hair pixels. The respective object mattes may be used to determine which pixels to adjust when adjusting pixel values such as to change color, lighting, texture, etc. The CNN may comprise a (pre-trained) network for image classification adapted to produce the segmentation mask. The CNN may be trained for image segmentation (e.g. using coarse segmentation data) to minimize a mask-image gradient consistency loss. The CNN may further use skip connections between corresponding layers of an encoder stage and a decoder stage where shallower layers in the encoder, which contain high-res but weak features are combined with low resolution but powerful features from deeper decoder layers.

Abstract: A method for compressing a neural network model includes acquiring a to-be-compressed neural network model. A first bit width, a second bit width and a target thinning rate corresponding to the to-be-compressed neural network model are determined. A target value is obtained according to the first bit width, the second bit width and the target thinning rate. Then the to-be-compressed neural network model is compressed using the target value, the first bit width and the second bit width to obtain a compression result of the to-be-compressed neural network model.

Abstract: This application provides a method, a terminal-side device, and a cloud-side device for data processing and a terminal-cloud collaboration system. The method includes: sending, by the terminal-side device, a request message to the cloud-side device; receiving, by the terminal-side device, a second neural network model that is obtained by compressing a first neural network model and that is sent by the cloud-side device, where the first neural network model is a neural network model on the cloud-side device that is used to process the cognitive computing task, and a hardware resource required when the second neural network model runs on the terminal-side device is within an available hardware resource capability range of the terminal-side device; and processing, by the terminal-side device, the cognitive computing task based on the second neural network model.

Abstract: A meta-learning system includes an inner function computation module, adapted to compute output data from applied input data according to an inner model function, depending on model parameters; an error computation module, adapted to compute errors indicating mismatches between the computed output data and target values; a state update module, adapted to update the model parameters of the inner model function according to an updated state, updated based on a current state of the state update module, in response to an error received from the error computation module. The state update module is learned to adjust the model parameters of the inner model function, such that a following training of the inner model function with training data is improved.

Abstract: A semantic segmentation method and apparatus for a three-dimensional image, and a storage medium are provided. The method includes: obtaining a three-dimensional image; slicing the three-dimensional image according to three directional planes, to obtain two-dimensional slice images of an x axis, two-dimensional slice images of a y axis, and two-dimensional slice images of a z axis; invoking a first segmentation model, a second segmentation model, and a third segmentation model to respectively perform semantic segmentation on the two-dimensional slice images of the x axis, the y axis, and the z axis, to obtain distribution probability maps of a target object on the three directional planes; and obtaining a three-dimensional distribution binary image of the target object by invoking an adaptive fusion model to perform three-dimensional fusion on the three distribution probability maps respectively corresponding to an x-axis directional plane, a y-axis directional plane, and a z-axis directional plane.

Abstract: A model receives a target demand curve as an input and outputs an optimized control sequence that allows equipment within a physical space to be run optimally. A thermodynamic model is created that represents equipment within the physical space, with the equipment being laid out as nodes within the model according to the equipment flow in the physical space. The equipment activation functions comprise equations that mimic equipment operation. Values flow between the nodes similarly to how states flow between the actual equipment. The model is run such that a control sequence is used as input into the neural network; the neural network outputs a demand curve which is then checked against the target demand curve. Machine learning methods are then used to determine a new control sequence. The model is run until a goal state is reached.

Abstract: Embodiments relate to system for automatically predicting payer response to claims. In an embodiment, the system receives claim data associated with a claim. The system identifies a set of claim features of the claim data, and generates an input vector with at least a portion of the set of claim features. The system applies the input vector to a trained model. A first portion of the neural network is configured to generate an embedding representing the input vector with a lower dimensionality than the input vector. A second portion of the neural network is configured to generate a prediction of whether the claim will be denied based on the embedding. The system provides the prediction for display on a user interface of a user device. The prediction may further include denial reason codes and a response date estimation to indicate if, when, and why a claim will be denied.

Abstract: A method of performing a class incremental learning in a neural network apparatus, the method including training an autoencoder using first input embeddings with respect to a first class group, calculating a contribution value of each of parameters of the autoencoder and calculating a representative value with respect to each of at least one first class included in the first class group in the training of the autoencoder, retraining the autoencoder using second input embeddings with respect to a second class group, and updating the contribution value of the each of the parameters and calculating a representative value with respect to each of at least one second class included in the second class group in the retraining the autoencoder.

Abstract: The disclosure discloses a method of executing dynamic graph for neural network computation and the apparatus thereof. The method of executing dynamic graph includes the following steps: S1: constructing and distributing an operator and a tensor; S2: deducing an operator executing process by an operator interpreter; S3: constructing an instruction of a virtual machine at runtime by the operator interpreter; S4: sending the instruction to the virtual machine at runtime by the operator interpreter; S5: scheduling the instruction by the virtual machine; and S6: releasing an executed instruction by the virtual machine. According to the method of executing dynamic graph for neural network computation and the apparatus thereof provided by the disclosure, runtime is abstracted to be the virtual machine, and the virtual machine acquires a sub-graph of each step constructed by a user in real time through the interpreter and schedules, the virtual machines issues, and executes each sub-graph.

Abstract: This disclosure relates generally to automated packing of objects, and, more particularly, to a method and system for packing products with increased efficiency across packaging levels. While conventional methods of improving packaging efficiency focus on only one of the multiple levels in the packaging process, most commonly the tertiary level, the present disclosure attempts increasing packaging efficiency across packaging levels. Embodiments of present disclosure achieves increased efficiency across packaging levels by identifying standard size of secondary packages for packing a plurality of primary packages, packing the secondary packages within tertiary packages using a Mixed Integer Linear Programming (MILP) optimization model based on packing heuristics, and providing a feedback between tertiary and secondary packaging levels to identify standard secondary packages which can pack the primary packages with higher packing efficiency.

Abstract: A computer-implemented content suggestion engine provides content suggestions to a requesting user based on information about content items that other users may have independently categorized or organized into folders within a content repository. Embodiments of the method comprise a content repository having a plurality of content items, where each content item is associated with one or more user-created folders. Embodiments further comprise receiving, via a network, a suggestion request for suggested content, where the suggestion request identifies a first content item for which suggestions are sought. Other content items in the content repository are then identified as potential suggestions based on the application of a formal relationship between the first content item and the potential suggested content items. One or more of the potential suggested content items may then be provided in response to the suggestion request via the network.

Abstract: Embodiments of the present disclosure relate to methods, systems and computer program products for causal analysis. In some embodiments, there is provided a computer-implemented method. The method comprises determining, from observation samples of a plurality of factors, a first causal structure indicating a first causal relationship among the plurality of factors, each observation sample including a set of observation values of the plurality of factors; presenting the first causal structure to a user; in response to receiving at least one user input about the first causal structure from the user, executing actions associated with the at least one user input based on the first causal structure; and presenting a result of the execution of the actions to the user. In other embodiments, another method, systems and computer program products are provided.

Abstract: A system and method for performing root cause analysis for enforcement events is presented. The system can enable accurate detection of an enforcement event and identifies the root cause of such events. The system can enable accurate detection of the enforcement event and identifies the root cause of such events using an automation workflow engine. The system can perform root cause analysis based on at least one analysis model. The system can provide a user with an interface to monitor the enforcement event by collecting a list of data points characterizing the enforcement event, as well as analyze the data points to evaluate what is the root cause of the enforcement event.

Abstract: A computer system is provided that is programmed to select feature sets from a large number of features. Features for a set are selected based on metagradient information returned from a machine learning process that has been performed on an earlier selected feature set. The process can iterate until a selected feature set converges or otherwise meets or exceeds a given threshold.

Abstract: Systems and methods may ethically evaluate intelligent systems operating in a real-world environment. The systems and methods may generate a clone of the intelligent system, and test the clone in a simulation environment. If the clone passes the testing, the systems and methods may permit the intelligent system to continue operating in the real-world environment. If the clone fails the testing, the systems and methods may override the intelligent system, such as disabling the intelligent system and assuming control in the real-world environment. The systems and methods may be implemented at a hardware level of a data processing device to prevent interference with the systems and methods by the intelligent system.

Abstract: A request is received associated with reviewing content. As part of the request, one or more conditions are received and the content is analyzed to identify a first field of interest and a second field of interest. The first field of interest and the second field of interest represent fields of interest associated with the review of the content. At least one of the first field of interest or the second field of interest may not satisfy the one or more conditions and the content, or a portion thereof, may be sent for review.

Abstract: A computer-implemented method for detecting and monitoring bias in an application includes index training data and obtaining a plurality of correlation values of one or more features in the indexed training data with a target variable. For each of the one or more features, a first value and a favorable result and a second value along with the unfavorable result is calculated. An absolute value of a difference between the calculated first value and the calculated second value is calculated. A total sum of the calculated absolute value of the plurality of correlation values of the one of the one or more features is calculated.

Abstract: A computer system is configured to receive a dataset of image-derived features for a plurality of images, reduce the dimensionality of this dataset, identify clusters within the dimensionally-reduced dataset, and generate a visual representation of the datapoint of the dimensionally-reduced dataset as icons grouped by cluster. User input is received to apply user classification labels to the images for inclusion in a training dataset. A user interface is useable to present information to the user and receive information from the user to facilitate the application of user classification labels.

Abstract: Systems and methods are disclosed for determining a propensity of an entity to take a specified action. In accordance with one implementation, a method is provided for determining the propensity. The method includes, for example, accessing one or more data sources, the one or more data sources including information associated with the entity, forming a record associated with the entity by integrating the information from the one or more data sources, generating, based on the record, one or more features associated with the entity, processing the one or more features to determine the propensity of the entity to take the specified action, and outputting the propensity.

Abstract: Methods, systems and computer program products for associating geographical locations with annotations corresponding to content. In one method, a language model is developed. The language model is developed from the location information and the one or more annotations associated with content uploaded by users. The language model is based on the probabilistic distribution of locations over one or more annotations. Further, when a user provides one or more annotations, the system and the method may use the language model to identify one or more locations associated with the one or more annotations provided by the user. The language model predicts one or more geographical locations based on the probabilistic distribution of locations over the annotations.

Abstract: The present disclosure relates to activity monitoring systems and methods for gating whether or not steps should be counted in an observation window based on whether a decision tree concludes there are consecutive step activities (versus no activity or other activities) in the observation window. Particularly, certain aspects are directed to a method that includes obtaining acceleration data for an observation window of an accelerometer, inputting two or more characteristics of the acceleration data into a decision tree to determine activity occurring within the observation window, assigning a first class to the observation window when the determined activity is associated with consecutive steps, assigning a second class to the observation window when the determined activity is not associated with consecutive steps, and when the first class is assigned to the observation window, determining a step count for the observation window using frequency analysis.

Abstract: System derives training change factors for services provided for training product user, priority assigned to training service ticket initiated by training product user, times of service ticket interactions associated with training service ticket, and/or age of training service ticket, and also for times of states of training service ticket. System uses training service ticket and training change factors to train change-based machine-learning model to predict change-based training probability that training product user escalated service for training service ticket. System derives change factors for services provided for product user, priority assigned to service ticket initiated by product user, times of service ticket interactions associated with service ticket, and/or age of service ticket, and also for times of states of training service ticket.

Abstract: A system for generating a statistical model for fault diagnosis comprising at least one hardware processor, adapted to: extract a plurality of structured values, each associated with at least one of a plurality of semantic entities of a semantic model or at least one of a plurality of semantic relationships of the semantic model, from structured historical information organized in an identified structure and related to at least some of a plurality of historical events, the semantic model represents an ontology of an identified diagnosis domain, each of the plurality of semantic entities relates to at least one of a plurality of domain entities existing in the identified diagnosis domain, and each of the plurality of semantic relationships connects two of the plurality of semantic entities and represents a parent-child relationship therebetween; extract a plurality of unstructured values, each associated with at least one of the plurality of semantic entities.

Abstract: An agricultural monitoring system, apparatus and method(s) for providing crop-related forecasts by performing the steps of receiving seasonal image data from at least one source, where the seasonal image data is associated with at least one agricultural field, processing the seasonal image data using a Bayesian framework, where the Bayesian framework comprises one or more crop models configured to predict, based on the seasonal image data, one or more probabilities indicative of at least one crop state, updating at least one crop model of the Bayesian framework based on the one or more probabilities, and outputting a forecast of the at least one crop state based on the one or more probabilities.

Abstract: A computer-implemented method comprising: obtaining, by way of an input, input data relating to speech provided by a user; deriving one or more hypotheses for each of a plurality of user data fields from the input data; obtaining one or more reference values for each of the plurality of user data fields for each of one or more candidate users; calculating a score for at least one candidate user of the one or more candidate users, calculating the score comprising: calculating a plurality of user data field scores comprising, for each of the plurality of user data fields, a respective user data field score using the one or more hypotheses and the one or more reference values for the candidate user for the respective user data field; performing one or more fuzzy logic operations on the plurality of user data field scores; using the score for a candidate user of the one or more candidate users to perform a verification or identification process for the user.

Abstract: A computer-implemented logistics method of arranging delivering of items to recipients situated at different recipient locations, comprising: recording items received at a distribution centre, clustering the records according to location of the recipient, locating a hub position for each cluster, further clustering the records, creating an individual schedule for each agent with events and locations and timings for the events, and instructing the agent to deliver the items by providing individual schedule information.

Abstract: An approach provides sending captured Superbill image data and output data generated based on results of parsing the captured Superbill image data to an external system which manages Superbill data via a cloud system and a storage service. The cloud system creates parsing rule data for parsing a captured Superbill image in accordance with user operation at a client device. The cloud system obtains captured Superbill image data from the storage service, in response to receiving a notification indicating that the captured Superbill image data has been stored in the storage service. The cloud system parses the obtained image data based on the created parsing rule and generates output data based on results of parsing. The cloud system sends the generated output data and the obtained image data to the external system via the one or more computer networks.

Abstract: A method includes receiving, in a first networking platform, an electronic message directed from a first party to a workflow to a second party of the workflow. The method also includes identifying a document attached to the electronic message as relevant to the workflow, and identifying at least a portion of a text content in the electronic message as relevant to the workflow. The method also includes updating the workflow associated with the workflow based on the document attached to the electronic message, when the second party provides the input and storing the document attached to the electronic message in a database, as a new version of the workflow. A system and a non-transitory, computer-readable medium storing instructions to perform the above method are also provided.

Abstract: Systems and methods for authenticating access to multiple data stores substantially in real-time are disclosed. The system may include a server coupled to a network, a client device in communication with the server via the network and a plurality of data stores. The server may authenticate access to the data stores and forward information from those stores to the client device. An exemplary authentication method may include receipt of a request for access to data. Information concerning access to that data is stored and associated with an identifier assigned to a client device. If the identifier is found to correspond to the stored information during a future request for access to the store, access to that store is granted.

Abstract: Systems and methods are provided for generating a base visual score for each candidate image of a plurality of images received by a computing system, based on the scene type of each image. For each candidate image, the computing system multiplies the base visual score by a feature importance weight to generate a first visual score, adds respective scene type bonus points to the first visual score to generate a second visual score, and adds diversity scoring points to the second visual score to generate a final visual score for each candidate image. The computing system ranks the candidate images based on the final visual scores and provides a specified number of the top-ranked candidate images to be displayed on a display of the computing device.

Abstract: Implementations of various methods and systems of a network, GPS system, mobile computing devices, servers, forward commodity market servers, grouping software for hubs, transparent open access pricing systems, blockchain audit and safety methods and systems, virtual hub systems, algorithm methods for no arbitrage conditions in a simple easy to use graphical user interface format for mobile or virtual computing over various mediums which are connected via a network to transact and trade transportation seats or capacity units and resulting financial swap payment structures in airline transport, subway transport, train transport, automobile transport, autonomous vehicle transport, taxi transport, space transport, package freight transport, tractor trailer freight transport, cargo freight transport, container freight transport, virtual transport, underground transport, ship or sea transport, public transport, private transport or drone transport on a computer, mobile computer device, audio computer device, virtu

Abstract: Concepts and technologies are disclosed herein for an infringement detection system that obtains images of products and images of proprietary objects, and analyzes the images to make coarse matches. An image comparison engine may transform or augment the data for comparison for coarse and refined matching. The outputs of the image comparison engine are initial infringement predictions that are further evaluated using refined matching including shape fitting. The detection system outputs refined infringement predictions, which may be optionally confirmed as counterfeit based on various considerations (e.g., known inauthentic products, suspect sales history, text and image anomalies, etc.). Upon a refined match or confirmation, the detection system records potential infringements and associated metadata into a database to initiate an optional response action.