Abstract: A computer platform includes an artificial neural network (ANN) as well as a classifier. The ANN is configured, after a learning phase, to transform an input data vector into a discriminating feature vector having a smaller dimension. A user then generates, from a plurality of reference data vectors, the same plurality of reference feature vectors, which are encrypted in an encryption module using the public key of a homomorphic cryptosystem and stored in a reference database of the platform. When the user requests the classification of an input data vector, the ANN, or a copy thereof, provides the classifier with a corresponding discriminating feature vector (y). Distances from the vector to the different reference feature vectors are calculated in the homomorphic domain and the index of the reference feature vector closest to y, i.e. the identifier i0 of the class to which it belongs, is returned to the user.

Abstract: This application discloses a keyword generating method, an apparatus, a device and a storage medium, which relate to the field of natural language processing in the field of artificial intelligence. A specific implementation scheme includes: inputting a target text into a text processing model, obtaining a word sequence corresponding to the target text, and generating a semantic representation sequence corresponding to the word sequence; making prediction about each semantic representation vector in the semantic representation sequence respectively to obtain a prediction result; and if the prediction result indicates that a word corresponding to the semantic representation vector is capable of triggering a generation of a keyword, outputting the keyword based on the semantic representation vector and the prediction result. This method improves the accuracy of generating keywords.

Abstract: A text display method, a text display apparatus, an electronic device and a storage medium are disclosed. A real scene image and a to-be-displayed text are acquired, motion track data for texts is invoked, the to-be-displayed text is processed with a dynamic special effect, and the text which has been subject to the dynamic special effect processing is displayed on a real scene image, thus realizing a function of displaying a text with dynamic special effect in augmented reality display, making the text display effect more vivid. The display method can be widely used in various application scenarios to bring users a better visual and sensory experience.

Abstract: The disclosure discloses a motor driving system converter fault diagnosis method based on adaptive sparse filtering, and belongs to the field of driving system fault diagnosis. The disclosure applies an unsupervised learning algorithm to an application scene of converter fault diagnosis. Effective features are automatically extracted from original data, and the problem of manual feature design based on expert knowledge is solved. Meanwhile, in consideration of current fundamental period change caused by different rotation speed working conditions, rotation speed feedback is introduced, secondary sampling is carried out on current sampled at a constant frequency, it is ensured that the length of a signal input into the deep sparse filtering network is one fundamental wave period, redundant information is better removed from original data, the calculation burden is relieved, and the accuracy and rapidity of the diagnosis algorithm are improved to a certain extent.

Abstract: A method, system, and computer readable storage medium are provided for predicting behavior of a vehicle. More specifically, image data of the vehicle is obtained by generating class prediction data for the vehicle and then identifying one or more related points of interest. The method, system, and computer readable storage medium then predict the behavior of the vehicle based upon the one or more points of interest and the class prediction data.

Abstract: A specification of a property required to be upheld by a computerized machine learning system is obtained. A training data set corresponding to the property and inputs and outputs of the system is built. The system is trained on the training data set. Activity of the system is monitored before, during, and after the training. Based on the monitoring, performance of the system is evaluated to determine whether the system, once trained on the training data set, upholds the property.

Abstract: Techniques for eliminating bias in selectively edited videos are provided. A request to release a video capturing a public safety incident is received. The video is edited to create an edited video. At least one civilian score and at least one public safety official score based on the sentiment of the video is computed. At least one edited civilian score and at least one edited public safety official score based on the sentiment of the video is computed. A first score is computed based on a combination of the civilian score and public safety official score. A second score is computed based on a combination of the edited civilian score and edited public safety official score. The first and second score are compared to determine if a difference between the scores exceed a threshold. The edited video is released when the scores do not exceed the threshold.

Abstract: Various embodiments of the present invention provide methods, apparatuses, systems, computing devices, computing entities, and/or the like for facilitating efficient and effective execution of database management operations. For example, various embodiments of the present invention provide methods, apparatuses, systems, computing devices, computing entities, and/or the like for facilitating efficient and effective execution of database management operations using at least one of query-compliant hash databases, segmentation-based hashing models, and hash segmentation models.

Abstract: A system for monitoring an environment may include an input device for monitoring and capturing pattern-based states of a model of the environment. The system may also include a thalamobot embodied in at least a first processor in communication with the input device. The thalamobot may include at least one filter for monitoring captured data from the input device and for identifying at least one state change within the captured data. The system may also include at least one critic and/or at least one recognition system. The at least one filter forwards said at least one state change to the critic and/or recognition system. Novel schemes are introduced to allow processors to interconnect themselves into brain-like structures that contemplate both the environment and the model thereof, unifying disparate data into discoveries. The significance of such discoveries is recognized either through neural activation patterns or the topologies of interconnecting neural modules.

Abstract: A data processing system implements receiving a request for user-centric topic recommendations from a computing device of a user in response to a user input in an application indicating that the user is attempting to assign a tag to a first content item in the application; obtaining a first set of user-centric topic recommendations from a first topic datastore based on a relevance ranking assigned to each of the topic recommendations; providing the first set of user-centric topic recommendations to the computing device; causing the computing device to display the first set of user-centric topic recommendations; receiving one or more second requests for user-centric topic recommendations; obtaining one or more second sets of user-centric topic recommendations from the first topic datastore based on the query string and the relevance ranking; and causing the computing device of the user to display the one or more second sets of user-centric topic recommendations.

Abstract: An artificial intelligence system and a method for searching for an optimal model are provided. A method for searching for a learning mode of an artificial intelligence system includes receiving, by an operator included in a first node, first channels, deriving, by the operator included in the first node, first parameter weight indexes corresponding to weights of first parameters by calculating the first parameters corresponding to each of the received first channels with the received first channels, generating and outputting a second channel group by combining the first channel with the other channel, receiving, by an operator included in a second node, second channels included in the second channel group, and deriving, by the operator included in the second node, second parameter weight indexes corresponding to weights of second parameters by calculating the second parameters corresponding to the received second channels with the received second channels.

Abstract: Systems and methods are provided for receiving a query created by a user, receiving output data of at least one function to retrieve data related to the query and analyzing the output data of the at least one function to retrieve data related to the query. The systems and methods further provide for generating at least one dynamic knowledge graph associated with the output data of the at least one function, wherein the at least one dynamic knowledge graph comprises data from the output data of the at least one function and indicates relationships between the data, analyzing the at least one dynamic knowledge graph to determine data relevant to the query generated by the user, and generating a response to the query based on the data relevant in the at least one dynamic knowledge graph.

Abstract: A method, a system, and a computer program product for performing analysis of data to detect presence of malicious code are disclosed. Reduced dimensionality vectors are generated from a plurality of original dimensionality vectors representing features in a plurality of samples. The reduced dimensionality vectors have a lower dimensionality than an original dimensionality of the plurality of original dimensionality vectors. A first plurality of clusters is determined by applying a first clustering algorithm to the reduced dimensionality vectors. A second plurality of clusters is determined by applying a second clustering algorithm to one or more clusters in the first plurality of clusters using the original dimensionality. An exemplar for a cluster in the second plurality of clusters is added to a training set, which is used to train a machine learning model for identifying a file containing malicious code.

Abstract: The present disclosure provides a soft memristor for soft neuromorphic system including a substrate, a first electrode layer formed on the substrate, a metal diffusion barrier layer formed on the first electrode layer, a resistive switching material layer formed on the metal diffusion barrier layer, and a second electrode layer formed on the resistive switching material layer.

Abstract: A self-clocking (or self-oscillating) modulator in signal processing, similar to a ?? modulator, with particular application in the design of neural networks based on such modulators is described. A system of multiple self-clocking modulators and supporting structures may be configured to perform a calculation similar to that of an analog computer, such as a neural network, at lower power and smaller size than a digital implementation. Such a system constructed using the present approach does not require a sequential solution, but rather converges on a solution in one step; unlike the typical prior art, it thus requires no clock and operates asynchronously in a manner similar to a conventional analog computer. The self-clocking modulator can function as a neuron in a neural network, receiving a sum-of-products signal and generating an output stream like that of a ?? modulator that represents this sum-of-products, potentially also including an activation function and offset.

Abstract: Methods and systems are provided for generating interpretable user modeling system. The interpretable user modeling system can use an intent neural network to implement one or more tasks. The intent neural network can bridge a semantic gap between log data and human language by leveraging tutorial data to understand user logs in a semantically meaningful way. A memory unit of the intent neural network can capture information from the tutorial data. Such a memory unit can be queried to identify human readable sentences related to actions received by the intent neural network. The human readable sentences can be used to interpret the user log data in a semantically meaningful way.

Abstract: The present disclosure involves systems, software, and computer implemented methods for a communication-efficient secret shuffle protocol for encrypted data based on homomorphic encryption and oblivious transfer. A service provider and multiple clients participate in a secret shuffle protocol of randomly shuffling encrypted client-specific secret input values. The protocol includes generation and exchange of random numbers, random permutations and different blinding values, including use of an oblivious transfer mechanism. A last protocol step includes using homomorphism, for each client, to perform computations on intermediate encrypted data to homomorphically remove a first blinding value and a second blinding value, to generate a client-specific rerandomized encrypted secret input value. As a result, the client-specific rerandomized encrypted secret input values are generated in an order that is unmapped to an order of receipt, at the service provider, of the encrypted secret input values.

Abstract: The present disclosure involves systems, software, and computer implemented methods for a efficient distributed secret shuffle protocol for encrypted database entries using dependent shufflers. Each of multiple clients provides an encrypted client-specific secret input value. A subset of clients are shuffling clients who participate with a service provider in a secret shuffling of the encrypted client-specific secret input values. The protocol includes generation and exchange of random numbers, random permutations and different blinding values. A last protocol step includes using homomorphism, for each client, to perform computations on intermediate encrypted data to homomorphically remove a first blinding value and a second blinding value, to generate a client-specific rerandomized encrypted secret input value.

Abstract: A neuromorphic synapse array is provided which ensures that a neuron model as such McCulloch-Pitts is dependent on nonlinearity with a single polarity weight cell. The neuromorphic synapse array includes a plurality of synaptic array cells, a plurality of operation column arrays, and a reference column array. The synaptic array cells respectively have a single polarity synapse weight and are classified into operation synapse cells and reference synapse cells for shifting a product-sum of the operation synapse cells. The operation column arrays are defined by the operation synapse cells aligned in column of the array. The reference column array is defined by the reference synapse cells aligned in column of the array.

Abstract: The method of present disclosure relates to neural chip and optimizing operation of a neural chip. The method includes sensing current values of physical parameters indicating an environment. Sensed current values are stored in a memory unit. The memory unit also stores previously sensed values of physical parameters. The current values and the previously sensed values are compared by the neural chip. Based on the comparison, one or more actions are applied using the previously sensed values, for completing the task, if the current values and the previously sensed values are matched. In case there is no matching, the neural chip uses the current valises for applying the one or more actions. The neural chip learns from applying of the actions and updates itself accordingly.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for machine translation using neural networks. In some implementations, a text in one language is translated into a second language using a neural network model. The model can include an encoder neural network comprising a plurality of bidirectional recurrent neural network layers. The encoding vectors are processed using a multi-headed attention module configured to generate multiple attention context vectors for each encoding vector. A decoder neural network generates a sequence of decoder output vectors using the attention context vectors. The decoder output vectors can represent distributions over various language elements of the second language, allowing a translation of the text into the second language to be determined based on the sequence of decoder output vectors.

Abstract: A method includes performing, with at least one processing device, natural language understanding by iteratively (i) generating a semantic word and clause representation and (ii) generating a syntax. The generation of the semantic word and clause representation and the generation of the syntax occur iteratively such that (i) semantics are calculated from syntax by aggregating weights of syntactically-labeled context in which words or clauses appear and (ii) syntax is calculated from semantics by grouping common pairs of words or clauses with similar semantic relations, thereby producing a self-consistent coupled notion of syntax and semantics.

Abstract: A method for generating hypotheses in a corpus of data comprises selecting a form of ontology; coding the corpus of data based on the form of the ontology; generating ontology space based on coding results and the ontology; transforming the ontology space into a hypothesis space by grouping hypotheses; weighing hypotheses included in the hypothesis space; and applying a science-based optimization algorithm configured to model a science-based treatment of the weighted hypotheses.

Abstract: Provided is a method of providing a recognition result of an object, the method including: detecting an object included in an image; determining an image group corresponding to the detected object through matching between the image and image groups each grouped for a same object; and outputting information of the image group in a front direction of the object of the determined image group, estimated from the image, based on a photographing condition of the determined image group and a photographing condition of the image.

Abstract: In an example, an apparatus comprises a plurality of processing unit cores, a plurality of cache memory modules associated with the plurality of processing unit cores, and a machine learning model communicatively coupled to the plurality of processing unit cores, wherein the plurality of cache memory modules share cache coherency data with the machine learning model. Other embodiments are also disclosed and claimed.

Abstract: Systems and methods for identifying a new device in an internet of things (IoT) network comprising a plurality of devices. The method includes, by a processor: detecting that a new device has been added to the IoT network; receiving identification data corresponding to the new device; determining a first device identification using a deep learning algorithm; determining a second device identification using a clustering algorithm; determining a third device identification based on a reference table; determining a fourth device identification by analyzing one or more values included in the identification data. The method further includes selecting, at least one of the first device identification, the second device identification, the third device identification, or the fourth device identification, as a final device identification by performing a weighted selection.

Abstract: Technical solutions are described for configuring a synaptic array. An example computer implemented method includes selecting a first electronic circuit and a second electronic circuit from the synaptic array for executing a task. The method further includes connecting the first electronic circuit to the second electronic circuit to facilitate passage of electric current by forming a metallic protrusion to connect a first connector of the first electronic circuit and a second connector of the second electronic circuit.

Abstract: An apparatus includes multiple parallel computing cores, where each computing core is configured to perform one or more processing operations and generate input data. The apparatus also includes multiple sets of parallel coprocessors, where each computing core is associated with a different one of the sets of parallel coprocessors. The coprocessors in each set of parallel coprocessors are configured to process the input data and generate output data. Each of the computing cores is configured to generate additional input data based on the output data generated by the associated set of parallel coprocessors.

Abstract: The disclosed technology generally relates to detecting and identifying objects in digital images, and more particularly to detecting, identifying and/or tracking moving objects in video images using an artificial intelligence neural network configured for deep learning. In one aspect, a method comprises capturing a video input from a scene comprising one or more candidate moving objects using a video image-capturing device, where the video input comprises at least two temporally spaced image frames captured from the scene. The method additionally includes transforming the video input into one or more image pattern layers, where each of the image pattern layers comprises a pattern representing one of the candidate moving objects. The method additionally includes determining a probability of match between each of the image pattern layers and a stored image in a big data library.

Abstract: As disclosed, f-scores can be generated for apparel items. Training images are identified, where each training image is associated with a corresponding set of tags including information about a plurality of attributes. A first convolutional neural network (CNN) is trained based on the plurality of training images and a first attribute. The first CNN is iteratively refined by, for each respective attribute, removing a set of neurons from the first CNN and retraining the first CNN based on the training images and the respective attribute. Upon determining that the first CNN has been trained based on each of the attributes, one or more CNNs are generated based on the first CNN. An image is received, where the image depicts an apparel item. The image is processed using the one or more CNNs, and an f-score for the apparel item is determined based on the output.

Abstract: A method comprising calculating output values of a plurality of linear nodes connected to a maxout node in a neural network, calculating a temporary maximum value among the output values during the calculation of the output values, and terminating the calculation of a final output value of a first linear node of the plurality of linear nodes in response to a condition that a difference between the temporary maximum value and a temporary output value of the first linear node exceeds a threshold value during the calculation of the output values.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for selecting answers to questions about documents. One of the methods includes receiving a document comprising a plurality of document tokens; receiving a question associated with the document, the question comprising a plurality of question tokens; processing the document tokens and the question tokens using a reader neural network to generate a joint numeric representation of the document and the question; and selecting, from the plurality of document tokens, an answer to the question using the joint numeric representation of the document and the question.

Abstract: Methods and apparatus, including computer program products, are provided for receiving, at a bidirectional recurrent neural network, a music file preprocessed to include at least one token data inserted within at least one location in the music file in order to enable varying the music file; generating, by the bidirectional recurrent neural network, an output music file, wherein the bidirectional recurrent neural network generates music data to replace the at least one token data; and providing, by the bidirectional recurrent neural network, the output music file representing a varied version of the music file. Related apparatus, systems, methods, and articles are also described.

Abstract: A method for modeling an operation of a system that may include a disturbance rejection model that is configured to generate a predicted value for a system output at a future time. The disturbance rejection model may include a neural network for mapping system inputs to the system output. The method may include the steps of: training the disturbance rejection model per a training dataset; and calculating a confidence metric for the disturbance rejection model. The confidence metric is configured to indicate a probability that a predicted sign of a gain in the system output at the future time made by the disturbance rejection model is correct.

Abstract: A method for generating hypotheses in a corpus of data comprises selecting a form of ontology; coding the corpus of data based on the form of the ontology; generating ontology space based on coding results and the ontology; transforming the ontology space into a hypothesis space by grouping hypotheses; weighing hypotheses included in the hypothesis space; and applying a science-based sorting algorithm configured to model a science-based treatment of the weighted hypotheses.

Abstract: A system for monitoring an environment may include an input device for monitoring and capturing pattern-based states of a model of the environment. The system may also include a 5 thalamobot embodied in at least a first processor, in which the first processor is in communication with the input device. The thalamobot may include at least one filter for monitoring captured data from the input device and for identifying at least one state change within the captured data. The system may also include at least one critic and/or at least one recognition system.

Abstract: A method and system for performing semantic analysis for electronic communication using a lexicon is provided. A neural network model is trained with a plurality of annotated text strings, the annotations comprising characteristic tuples that indicate characteristics for the text strings. An unannotated text string is received that comprises a plurality of words from a user. A characteristic matrix for the received text string is generated using a lexicon. The determined characteristic matrix is input into the trained neural network. And a characteristic tuple that indicates a characteristic for the received text string is received as output from the trained neural network.

Abstract: A machine learning system according to an embodiment of the present invention includes a state observer for observing the rotation number, torque, current, and voltage values of a motor detected by a motor controller for driving the motor; a motor output calculator for calculating a motor output from the rotation number, torque, current, and voltage values of the motor observed by the state observer; a reward calculator for calculating a reward based on the motor output; and a learning unit for updating an action value table based on the rotation number, torque, current, and voltage values of the motor.

Abstract: A computing system includes a programming interface of a control interface of a distributed computing environment, a service layer of the control interface, and a manager of the control interface. The programming interface is configured to receive a block of a block chain database. The block includes a ledger that includes a plurality of transactional data records. The service layer is configured to analyze the plurality of records to determine that one of the plurality of records is an indication of a request by a client for a service provided by a data interface of the distributed computing environment. The manager is configured to allocate access to execute the request in response to receiving the indication of the request.

Abstract: Long short term memory units that accept a non-predefined number of inputs are used to provide natural language relation extraction over a user-specified range on content. Content written for human consumption is parsed with distant supervision in segments (e.g., sentences, paragraphs, chapters) to determine relationships between various words within and between those segments.

Abstract: A machine learning device which learns the cleaning frequency of a fan motor which cools an electric motor mounted in a machine includes a state observation unit which observes the state of the machine, and a learning unit which updates an action value table for cleaning the fan motor, on the basis of a state variable observed by the state observation unit. This can improve both the life of the electric motor and the operating ratio of the machine.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for predicting object identity using an ensemble of predictors. In one aspect, a method includes selecting candidate objects that likely match a received object that is to be identified, from a database of objects, and providing attributes of the received object compared with those of the candidates to an ensemble of predictors having respective properties. Based on previous training, each predictor can predict a most likely candidate. From among the most likely candidates, a previously trained support vector machine can select a potential match candidate. If a score that the support vector machine associates with the potential match candidate, that is representative of the potential match candidate's likelihood to match the received candidate satisfies a threshold, then the potential match candidate can be determined to be the received candidate.

Abstract: Embodiments of the present invention provide a memristor having a first electrode, a second electrode and a memristive layer arranged between the first electrode and the second electrode. Thereby, the memristor is adapted to obtain an asymmetrical current density distribution in the memristive layer.

Abstract: Systems and methods are provided for distributing a playlist of a user. The systems and methods receive an indication from the user of a song to add to the playlist and adds the song to the playlist. In aspects, the systems and methods analyze the playlist to determine whether an attribute of the playlist meets a specified parameter. In some cases, the specified parameter can be met if a musical attribute of the song is similar to a music profile of one of the users, or if an amount of songs on the playlist meets or exceeds a specified threshold. The systems and methods can enable the user to distribute the playlist either publicly or to a specified set of users if the specified parameter of the playlist is met.

Abstract: Methods of animating objects using the human body are described. In an embodiment, a deformation graph is generated from a mesh which describes the object. Tracked skeleton data is received which is generated from sensor data and the tracked skeleton is then embedded in the graph. Subsequent motion which is captured by the sensor result in motion of the tracked skeleton and this motion is used to define transformations on the deformation graph. The transformations are then applied to the mesh to generate an animation of the object which corresponds to the captured motion. In various examples, the mesh is generated by scanning an object and the deformation graph is generated using orientation-aware sampling such that nodes can be placed close together within the deformation graph where there are sharp corners or other features with high curvature in the object.

Abstract: A communication system uses a bus to transmit information, by receiving signals and mapping them to a second set of signals representing codewords of a superposition signaling code, and transmitting the second set of signals. The superposition signaling code can comprise more than one layer. The pin-efficiency can be larger than 1. The system may encode bits into a codeword of a superposition signaling code that is defined by two basis vectors of predetermined size and then have two encoders for permutation modulation codes defined by the basis vectors. The bits of information are divided into a first part representing a predetermined number of bits and a second part representing a predetermined number of bits, with the parts provided to the respective encoding circuits and their outputs combined by a superposition.

Abstract: Described is a system for allowing sets of processors to engage in a secure pattern matching protocol. An input pattern is received from a first set of processors, while a text is received from a second set of processors. A matrix is constructed based on values computed for each character determined by each character's position in the pattern. The first set of processors sends an encrypted matrix to the second set of processors. The second set of processors processes each character in the text and creates a set of vectors. A final activation vector is created based on processing the set of vectors and an encrypted activation vector. The second set of processors sends the final activation vector to the first set of processors. The second set of processors decrypts the final activation vector. The system provides to the first set of processors where the pattern matches the text.

Abstract: Systems and methods for performing predictive analytics in an electrical grid network are disclosed. In one example of the disclosed technology, a method comprises aggregating a plurality of events from an electrical grid network, analyzing the plurality of events to recognize at least one event pattern, serializing at least one of the event patterns in a database, and predicting a future event pattern based on a correlation of the plurality of event patterns.

Abstract: Systems and methods for hierarchically partitioning a data set including a plurality of offerings are described. A system receives a data set including a plurality of offerings characterized by one or more offering attributes. The system identifies one or more candidate offering attributes associated with the offerings according to which the offerings are to be partitioned. The system assigns a partition hierarchy level to each of the candidate offering attributes that indicates a hierarchy of the offering attribute relative to other candidate offering attributes. The system determines, for each of the candidate offering attributes, a plurality of attribute values according to which the offerings are to be partitioned.

Abstract: A method for learning connections between nonlinear oscillators in a neural network comprising the steps of providing a plurality of nonlinear oscillators, with each respective oscillator producing an oscillation distinct from the others in response to an input and detecting an input at an at least first oscillator of the plurality of nonlinear oscillators. Detecting an input at an at least a second oscillator of the plurality of nonlinear oscillators, comparing the oscillation of the at least first oscillator to the oscillation of the at least second oscillator at a point in time, and determining whether there is coherency between the oscillation of the at least first oscillator and the oscillation of the at least second oscillator. Changing at least one of the amplitude and phase of a connection between the at least first oscillator and the at least second least oscillator as a function coherency between the at least first oscillator and the oscillation of the at least second oscillator.