Abstract: According to one embodiment, a learning device includes an acquirer, a generator, a first identifier, an extractor, a second identifier, and a learning-processor. The acquirer acquires real data. The generator generates pseudo data of the real data using a first neural network. The first identifier identifies whether input data is the real data or the pseudo data. The extractor extracts features of data from the input data. The second identifier identifies whether the features extracted by the extractor are features of the real data or features of the pseudo data. The learning-processor learns the first neural network such that the pseudo data and the real data are not discriminated on the basis of an identification result of each identifier.

Abstract: Methods and systems for detecting and classifying defects on a specimen are provided. One system includes one or more components executed by one or more computer subsystems. The one or more components include a neural network configured for detecting defects on a specimen and classifying the defects detected on the specimen. The neural network includes a first portion configured for determining features of images of the specimen generated by an imaging subsystem. The neural network also includes a second portion configured for detecting defects on the specimen based on the determined features of the images and classifying the defects detected on the specimen based on the determined features of the images.

Abstract: Embodiments of the present systems and methods may provide the capability to classify medical images, such as mammograms, in an automated manner using existing annotation information. In embodiments, only the global, image level tag may be needed to classify a mammogram into certain types, without fine annotation of the findings in the image. In an embodiment, a computer-implemented method for classifying medical images may comprise receiving a plurality of image tiles, each image tile including a portion of a whole view, processed by a trained or a pre-trained model and outputting a one-dimensional feature vector for each tile to generate a three-dimensional feature volume and classifying the larger image by a trained model based on the generated three-dimensional feature volume to form a classification of the image.

Abstract: A system includes a model repository comprising a plurality of models respectively being adapted to perform, when used by an analytical program, a computational task, in which a first database table is created in the database, the first database table having a predefined table structure that corresponds to the analytical program, a best-model of the plurality of models is stored in the first database table, and a request of a client device to perform the computational task and comprising input data is received. If the received request does not comprise a model-ID, the analytical program reads the model currently stored in the first table and uses the read model for performing the computational task on the input data.

Abstract: Systems and methods for performing OCR of an image depicting text symbols. An example method comprises: receiving an original image of a document; identifying location of a symbol sequence in the image; performing a series of conversion operations on a the portion of the image containing the identified symbol sequence; performing OCR of the symbol sequence in the resulting images; and combining these OCR results to produce a resulting OCR text representing at least a portion of the original document.

Abstract: A method for transmitting a parameter of a storage chip of an imaging cartridge, a storage chip and an imaging cartridge are provided. The method includes: receiving an instruction transmitted by a print imaging device; and determining that the instruction is an instruction for reading a key parameter, and transmitting the key parameter to the print imaging device, where the key parameter is generated by the storage chip. In the application, the parameter to be read by the print imaging device is generated by the storage chip. When the printing imaging equipment transmits the reading instruction, the storage chip transmits the parameter generated by itself to the print imaging device. Therefore, the parameter cannot be directly read by the printing imaging equipment at any moment, and thus a certain important parameter in the storage chip can be prevented from being obtained maliciously.

Abstract: The disclosure discloses a printer comprising a feeder, a printing head, a printer RFID tag comprising a tag antenna and an IC circuit part, a communication antenna, and a control device. The control device being configured to execute an information acquisition process for acquiring access point information via a wired communication connecting part from the printer RFID tag, the access point information being acquired by the printer RFID tag by tag communication, a transmission and reception attempting process for attempting information transmission and reception with an access point, by means of the mutually recognized wireless communication, a data acquisition process for acquiring desired print data acquired by the mutually recognized wireless communication when information transmission and reception by the transmission and reception attempting process succeed, and a printing control process for controlling the feeder and the printing head so that printing is performed on a print-receiving medium.

Abstract: A barcoded indicator operative to provide a machine-readable indication of exceedance of at least one threshold by at least one product quality affecting parameter, the barcoded indicator including at least a first barcode and at least a second barcode, the at least a second barcode being in a second barcode unreadable state prior to exceedance of the at least one threshold and upon exceedance of the at least one threshold the at least a first barcode becoming unreadable and generally simultaneously the at least a second barcode becoming readable.

Abstract: A transaction card is provided. The transaction card includes a card frame having a card inlay and a card housing. The transaction card also includes a magnetic stripe disposed inside the card frame between the card inlay and the card housing.

Abstract: Some embodiments relate to a device for checking the presence of articles on a shelf, that includes a detector circuit for detecting the presence of at least one article placed in at least one area of the shelf, the detector circuit being capable of switching between an article detection state, wherein it detects the presence of at least one article in the area, and an article non-detection state, wherein it does not detect any article in the area, and an RFID tag connected to the detector circuit and capable of reading and storing the article detection state or article non-detection state, of the detector circuit, the RFID tag further including an identification code for identifying the shelf area and being arranged to communicate the identification code and the state of the detector circuit to a remote RFID reader.

Abstract: Systems and methods for indicating tampering of a stored value card. The systems generally include a stored value card having a first activation indicia, and a removable label having a second activation indicia and covering the first activation indicia. To activate the stored value card, the first and second activation indicia are scanned. The removable label is configured to render the second activation indicia unreadable and/or to provide visual evidence of tampering such that activation is not recommended.

Abstract: A transaction card and a process of making the transaction card are described. The transaction card includes a core having first and second faces, a core thickness therebetween, and an opening, and embedded electronics disposed in the opening.

Abstract: An RFID tag includes a base material, an RFIC element, and a dipole antenna. The base material has a first end, a second end, a first side, and a second side. The RFIC element is mounted on the base material and includes a first input and output terminal and a second input and output terminal. The dipole antenna includes a first dipole element and a second dipole element. The first dipole element is a conductor pattern that extends from a first connection end toward the first end and meanders toward the first side, and the second dipole element is a conductor pattern that extends from a second connection end toward the second end and meanders toward the second side.

Abstract: A computer system is optimized for implementing a neural network nodal graph that has dense inputs and sparse inputs. The computer system has a local machine that receives user inputs and is optimized for computing power, and has a remote machine that stores embedding matrices and parameters, and is optimized for memory capacity. In accordance with a cost function applied to each node, the neural network nodal graph is divided into graph segments based on its types of inputs and needed computing resources for execution. In accordance with the cost functions, the graph segments are divided between the remote and local machines for execution, and the results of all the graph segments are combined in the local machine.

Abstract: A preprocessing module of a neural network has a first input and second input. The module generates multiple, different first latent vector representations of its first input, and multiple, different second latent vector representations of its second input. The module then models pairwise interactions between every unique pairwise combination of the first and second latent vector representations. The module then produces an intermediate output by combining the results of the modeled pairwise interactions.

Abstract: An apparatus and method for local quantization for convolutional neural networks. For example, one embodiment of an apparatus comprises: a convolutional neural network module comprising a neuron network structure to perform pattern recognition within an input image using a set of input image values; and a quantization module to quantize input image values to reduce processing requirements within one or more stages of the neuron network structure; the quantization module to perform quantization of each of a plurality of patches of the input image using a first quantization policy to generate a first matrix of quantized input data and to perform quantization of each of a plurality of kernel data using a second quantization policy to generate a second matrix of quantized kernel data.

Abstract: Aspects for forward propagation of a convolutional artificial neural network are described herein. The aspects may include a direct memory access unit configured to receive input data from a storage device and a master computation module configured to select one or more portions of the input data based on a predetermined convolution window. Further, the aspects may include one or more slave computation modules respectively configured to convolute a convolution kernel with one of the one or more portions of the input data to generate a slave output value. Further still, the aspects may include an interconnection unit configured to combine the one or more slave output values into one or more intermediate result vectors, wherein the master computation module is further configured to merge the one or more intermediate result vectors into a merged intermediate vector.

Abstract: Aspects for forward propagation of a multilayer neural network (MNN) in a neural network processor are described herein. As an example, the aspects may include a computation module that includes a master computation module and one or more slave computation modules. The master computation module may be configured to receive one or more groups of MNN data. The one or more groups of MNN data may include input data and one or more weight values and wherein at least a portion of the input data and the weight values are stored as discrete values. The one or more slave computation modules may be configured to calculate one or more groups of slave output values based on a data type of each of the one or more groups of MNN data.

Abstract: A three-dimensional (3D) ultra-low power neuromorphic accelerator is described. The 3D ultra-low power neuromorphic accelerator includes a power manager as well as multiple tiers. The 3D ultra-low power neuromorphic accelerator also includes multiple cores defined on each tier and coupled to the power manager. Each core includes at least a processing element, a non-volatile memory, and a communications module.

Abstract: In one embodiment, a method includes a preprocessing stage of a neural network model, where the preprocessing stage includes first and second preprocessing modules. Each of the two modules has first input that may receive a dense input and a second input that may receive a sparse input. Each module generates latent vector representations of their respective first and second inputs, and combine the latent vectors with the original first input to define an intermediate output. The intermediate output of the first module is fed into the first input of the second module.

Abstract: According to an embodiment, a learning device includes a calculator and a learner. The calculator is configured to calculate a value of a first objective function and a value of a second objective function. The first objective function includes smoothness that indicates smoothness of a local distribution of an output of a model, and is used to estimate a first model parameter for determining the model. The second objective function is used to estimate, with a second model parameter that is a hyperparameter of a learning method of learning, the model by using the first objective function. The second model parameter to be estimated is closer to a distance scale of learning data. The learner is configured to update the first model parameter and the second model parameter so that the value of the first objective function and the value of the second objective function are optimized.

Abstract: A learning apparatus includes a first learning unit, a learning data generator, and a second learning unit. The first learning unit implements a learning process for a first classifier such that a class applicable degree of a candidate class corresponding to a correct answer class becomes the maximum as compared with class applicable degrees of other candidate classes. The learning data generator performs a classification process for a classification target using the first classifier already subjected to the learning process, converts class applicable degrees of a plurality of candidate classes, which are output by the first classifier already subjected to the learning process into a dimensionally compressed value based on a predetermined compression rule, and generates second learning data in which the identification target is associated with the dimensionally compressed value. The second learning unit implements a learning process for a second classifier using the second learning data.

Abstract: A computer architecture for processing an aggregate dataset in an artificial neural network includes a master processor having a primary detector configured to analyze the aggregate dataset and segregate the aggregate dataset into component datasets, and two or more processing nodes in communication with the master processor, each of the processing nodes having secondary detectors configured to analyze the component datasets, wherein the master processor assigns the component datasets to the processing nodes based on processing capabilities of the processing nodes, and wherein the secondary detectors identify data labels associated with the processing nodes by analyzing the component datasets.

Abstract: An optimized computer architecture for training an neural network includes a system having multiple GPUs. The neural network may be divided into separate portions, and a different portion is assigned to each of the multiple GPUs. Within each GPU, its portion is further divided across multiple training worker threads in multiple processing cores, and each processing core has lock-free access to a local parameter memory. The local parameter memory of each GPU is separately, and individually, synchronized with a remote master parameter memory by lock memory access. Each GPU has a separate set of communication worker threads dedicated to data transfer between the GPU and the remote parameter memory so that the GPU's training worker threads are not involved with cross GPU communications.

Abstract: A method includes generating, by a processor of a computing device, a first plurality of models (including a first number of models) based on a genetic algorithm and corresponding to a first epoch of the genetic algorithm. The method includes determining whether to modify an epoch size for the genetic algorithm during a second epoch of the genetic algorithm based on a convergence metric associated with at least one epoch that is prior to the second epoch. The second epoch is subsequent to the first epoch. The method further includes, based on determining to modify the epoch size, generating a second plurality of models (including a second number of models that is different than the first number) based on the genetic algorithm and corresponding to the second epoch. Each model of the first plurality of models and the second plurality of models includes data representative of neural networks.

Abstract: A content review system for an online system automatically determines if received content items to be displayed to users violate any policies of the online system. The content review system generates a semantic vector representing the semantic features of a content item, for example, using a neural network. By comparing the semantic vector for the content item with semantic vectors of content items previously determined to violate one or more policies, the content review system determines whether the content item also violates one or more policies. The content review system may also maintain templates corresponding to portions of semantic vectors shared by multiple content items. An analysis of historical content items that conform to the template is performed to determine a probability that received content items that conform to the template violate a policy.

Abstract: A content review system for an online system automatically determines if received content items to be displayed to users violate any policies of the online system. The content review system generates a semantic vector representing the semantic features of a content item, for example, using a neural network. By comparing the semantic vector for the content item with semantic vectors of content items previously determined to violate one or more policies, the content review system determines whether the content item also violates one or more policies. The content review system may also maintain templates corresponding to portions of semantic vectors shared by multiple content items. An analysis of historical content items that conform to the template is performed to determine a probability that received content items that conform to the template violate a policy.

Abstract: A system for applications including but not limited to machine learning and unsupervised learning for clustering of data processes an input dataset in a plurality of autoencoders where each autoencoder is configured to produce an indication of a particular structure included in the input dataset and an aggregator combines the indications produced by the plurality of autoencoders based on a weighting vector to produce a weighted combination of the indications that may be used to train the system to create a sparse representation of unlabeled datasets.

Abstract: A Mechanical Turk—Integrated Development Environment system is disclosed. An integrated development environment (IDE) can include one or more interfaces capable of communicating with a mechanical turk engine. As a developer creates applications within the IDE, the developer can use the IDE to submit one or more requests to the mechanical turk engine. The engine constructs a mechanical turk project based on the requests and provides project tasks to workers. The results of the tasks can then be compiled and integrated back into the developer's application via the IDE. An example use includes constructing large domain specific data sets that can be applied to spoken dialog interfaces.

Abstract: Certain example embodiments described herein relate to techniques for determining the effectiveness of social media content posted to a particular network location. An example technique includes receiving a set of social media content records posted to a network location and one or more user reaction records posted in response to the social media content record, associating at least one emotion token with each received user reaction record, assigning at least one emotion category to each social media content record, obtaining one or more engagement metrics for each social media content record, and determining a relationship between at least one engagement metric and at least one emotion category based upon the obtained one or more engagement metrics and the assigned at least one emotion category of respective social media content records in the set.

Abstract: Systems and methods are provided for, are provided for analyzing each of a plurality of documents related to a person to identify content attributes that occur in a dictionary for behavior or personality traits, calculating an initial score for each behavior and personality trait related to the identified content attributes, calculating trait metrics for each behavior or personality trait by combining initial scores for each behavior or personality trait, providing the trait metrics as input to a rule based scoring and machine learning system, obtaining, from the rule based scoring and machine learning system, a trustworthiness score of the person output from the rule based scoring and machine learning system, receiving an indication that a second person trusts the person, creating a relationship between the second person and the person, and adjusting the trustworthiness score of the person based on, at least, a trustworthiness score of the second person.

Abstract: A method includes receiving a first query by a computing device and assigning the first query to a plurality of cognitive engines, wherein each of the plurality of cognitive engines include different characteristics for processing data. The method also includes, responsive to receiving a response from each of the plurality of cognitive engines for the first query, comparing the received responses from the plurality of cognitive engines. The method also included responsive to determining a difference between a first response from a first cognitive engine and a second response from a second cognitive engine is above a predetermined threshold value, performing a response mediation process until the difference is below the predetermined threshold value. The method also includes selecting a first final response from the received responses for the first query and the second query and displaying the first final response to a user.

Abstract: A system includes one or more memory devices storing instructions, and one or more processors configured to execute the instructions to perform steps of a method for generating a data tree structure for expediting rule-based decision-making. The system may receive a rules file having a plurality of rules, where each rule has one or more conditions and one or more actions. The system may generate a plurality of condition frequency counts that represent the total number of rules having a condition of a particular unique condition type. The system may then generate a tree structure based on the plurality of rules and the plurality of condition frequency counts, where the tree structure includes a root node and one or more node branches, where each node branch includes a first-level node and zero or more second-level nodes that each include a node condition and zero or more node actions.

Abstract: Methods and system are disclosed that bypass a maintenance mode of an application, for some users. In one aspect, multiple requests may be received to establishing a connection with multiple applications at a load balancer component. Upon receiving the received requests, a request processing and inspection unit and an intelligent control unit may be instantiated. The request processing and inspection unit may preprocess or process the received requests. The preprocessed or processed requests may be forward to the intelligent control unit. Based on a determination of header information and a content of the received requests, the intelligent control unit may determine whether to execute a routine to bypass a maintenance mode of an application and direct the requests to establish connection with the application or to execute a routine related to the maintenance mode of the application.

Abstract: Embodiments for recommending meals by a processor. A collaboration of data capturing a plurality of factors of a group user profile for each user in a group of users may be received for aiding in recommending one or more meals. The one or more meals may be recommended for the group of users according to the group user profile such that the recommending balances a satisfaction level for the one or more meals for the group of users.

Abstract: Methods and systems consistent with the present disclosure identify useful ways or means of comparing and contrasting information from automated systems as compared to human inputs when identifying potential weaknesses in designs or other methods. This disclosure is also associated with identifying apparatus/systems and methods that combine the skills of humans at sensing and understanding unusual situations with the capabilities of machines and artificial intelligence (AI) to create a better more reliable type of warning system for practical application when attempting to improve outcomes, especially in situations where a high consequence is associated with a low probability. These systems and methods may also be useful in conditions where a negative outcome is associated with a high probability and where new perspectives can identify or lead to solutions or warnings that may lead to a positive outcome when those new perspectives are associated with a low or unknown success probability.

Abstract: A framework system and method for developing hybrid voice/video QoE predictors, which use both network/codec/client parameters as well as the voice/video reference sample(s). The prediction model/algorithm uses deep packet inspection to produce relevant input and therefore the network's impact on the voice/video QoE can be determined without recording actual media (voice/video) content. Therefore, the QoE predictors are neither solely media based as available perceptual metrics, nor available parametric based. The present hybrid voice/video QoE uses the reference/original media information, unlike prior art hybrid video only QoE which use the recorded media (video only).

Abstract: A method for conversation branching may include storing a plurality of messages communicated in an online conversation and generating a fingerprint for each message. The fingerprint for each message may be stored in a lookup table. The method may also include detecting a new message from one of the users and processing the new message to generate an artifact that correlates to the new message. The lookup table may be queried using the artifact to determine a similarity between the new message and each of the stored messages. One or more proposed reply messages to the new message may be generated for branching the online conversation to a new conversation point that corresponds to a selected one of the one or more proposed reply messages. The one or more proposed reply messages are based on one of the stored messages that has a closest similarity to the new message.

Abstract: A biosignal acquisition system for transmitting context commands based on sensed biosignals is disclosed. The biosignal acquisition system may include a computer, a biosensor and a biosignal acquisition device. The biosensor may be configured to be in contact with a scalp of an individual. The biosignal acquisition device may be operably coupled to the biosensor and configured to amplify signals received by the biosensor and communicate the amplified signal to the computer. The computer may be configured to receive the amplified signals from the biosignal acquisition device, generate a feature matrix based on the amplified signals, identify a brain switch associated with each feature in the feature matrix, predict the probability that a classifier is associated with the brain switch of each feature, and fit the classifier based on the feature matrix.

Abstract: Disclosed embodiments include systems and methods relevant to optimization of dynamic content. For example, disclosed embodiments can involve generating dynamic content based on the performance of content variants. The performance of content variants can be analyzed and used as input to a machine learning framework that allows for the creation of variants configured to optimize for goal parameters. In some examples the machine learning framework can include the use of greedy algorithms, discrete hill climbing algorithms, and evolutionary algorithms among others.

Abstract: In one embodiment, a method includes a system building a first machine-learning module and one or more secondary machine-learning modules for operating with an application. The first and second modules may be configured to utilize, in operation, particular types of processing hardware, respectively. The system may receive from a client device a request to download the application, and in response send to the device the application with the first module. The system may then receive another request to download a selected one of the secondary modules, which may be selected based on a determination by the application running on the device that the associated type of processing hardware is available. In response, the system may send the selected module to the device. The application may be configured to selectively use the first module or the selected module to perform an operation.

Abstract: Some embodiments are directed to a computation system for learning a classifier from a data set and associated classification data. The system comprises multiple data devices and multiple computation devices. The data set and associated classification data is stored across the multiple data devices. Classifier parameters of the classifier are learned iteratively from the data set. The learning comprises computing local error terms for the classifier from the associated classification data and from classifier results obtained from applying the classifier with the current classifier parameters to the local part of the data set, and providing the local error term as a private input to the multiple computation devices for a multi-party computation among the multiple computation devices.

Abstract: An alarm handling system in plant process automation with a data processing device includes: at least one interface for accessing and/or processing one or more process signals and for determining corresponding process variables; an alarm configuration device for accessing and/or providing alarm configuration information including at least one setpoint for one or more determined process variables; and a prediction device for determining and processing a current rate of change of at least one process variable to predict how long it will take and/or a period until and/or predict at which date and/or time a provided setpoint and/or threshold is reached and/or crossed, and/or for determining whether and/or when at least one of the monitored and/or determined process variable values will cross the respective setpoint.

Abstract: Systems and methods for cross-session response tracking in a distributed computing environment using query identifiers generated based on search parameters. A search query including search parameters is received from a client device. A query identifier for the search query is generated based on the search parameters. One version of a search result among multiple versions of the search result is selected based on the query identifier and sent to the client device. A record in a transaction database is correlated with the selected version of the search result using the query identifier.

Abstract: The present invention provides a method, system, and computer program product of real-time cognitive supply chain optimization. In an embodiment, the present invention includes receiving supply chain optimization data describing a supply chain optimization problem, receiving variables of interest data describing variables of interest with respect to the supply chain optimization problem, generating an optimal route for a user based on the supply chain optimization data and the variables of interest data, monitoring route execution data describing execution of the generated optimal route by the user, capturing deviation data from the monitored route execution data, with respect to critical variable data, and executing a set of logical operations analyzing the captured deviation data with respect to the variables of interest data to identify whether at least one change occurred among the variables of interest at the time of at least one deviation from the generated optimal route.

Abstract: In accordance with embodiments, there are provided mechanisms and methods for progressive rendering of report results. These mechanisms and methods for progressive rendering of report results can enable embodiments to render report results in portions as they are received. The ability of embodiments to render report results in portions as they are received can enable report results to be progressively rendered such that delay is avoided Which would otherwise occur when rendering only in response to receipt of an entirety of the report results.

Abstract: Some embodiments of the invention provide systems and methods for deploying employees to roles in a restaurant. For example, some embodiments are directed to basing employee deployment decisions at least in part on a combination of quantitative data and qualitative assessments of performance by various employees in various roles. Qualitative assessments may, for example, relate to an employee's success in promoting customer satisfaction, or to the accuracy of quantitative data relating to a task or function.

Abstract: A device may receive data related to a material. The data may permit an analysis of an exposure related to the material. The data may be received from another device. The device may process the data to determine a category of materials for the material. The category of materials may indicate a set of exposures, related to the material, to an organization. The device may process the data to determine an exposure preparedness level associated with the material. The exposure preparedness level may indicate a preparedness of the organization to handle the set of exposures related to the material. The device may determine a score for the material based on the category of materials and the exposure preparedness level associated with the material. The score may indicate the exposure related to the material. The device may perform an action related to the material based on the score for the material.

Abstract: Embodiments are directed towards identifying and decreasing operational pain and increasing system efficiency through health monitoring and management. This may be accomplished through measuring, monitoring, reducing meaningful incident behavior across an organization and using it to inform necessary changes in the organizational operations to improve efficiency, or the like. Ergonomic data or metrics collected by a resource management engine may be used to intelligently inform management decisions to increase human well-being across the organization's workforce to optimize overall system performance. Accordingly, resource management engines may identify areas in organizations that need improvement or repair. In some embodiments, resource management engines may be arranged to participate in a continuous feedback loop that may provide continuous overall system optimization.

Abstract: Embodiments relate to systems and methods for customer interaction identification and analytics. A system can include a plurality of monitoring devices, a task management system and a customer interaction identification and analytics module. The plurality of monitoring devices can be configured for wear or handheld use and include a housing, a portable electronic computing device, and at least one sensor to sense activity data regarding a person wearing or holding the monitoring device. The task management system is communicatively coupled the plurality of monitoring devices to present at least one task to be completed in a retail environment, the at least one task associated with a temporal reference and a location in the retail environment. The customer interaction identification and analytics module, includes a gesture recognition database, communicatively coupled with the task management system and the plurality of monitoring devices.