Abstract: A data processing system for generating predictive maintenance models is disclosed, including one or more processors, a memory, and a plurality of instructions stored in the memory. The instructions are executable to receive a historical dataset relating to each system of a plurality of systems, the historical dataset including maintenance data and operational data. The instructions are further executable to evaluate correlation between each of a plurality of operational data features and maintenance events of the maintenance data, using a supervised machine learning classification model. The instructions are further executable to display a quantitative result of the evaluation for each operational data feature in a graphical user interface, receive a selection of one or more operational data features of the plurality of operational data features, and generate a predictive maintenance model using the selected one or more operational data features according to a machine learning method.

Abstract: Aspects of the disclosure relate to detecting and identifying malicious sites using machine learning. A computing platform may receive a uniform resource locator (URL). The computing platform may parse and/or tokenize the URL to reduce the URL into a plurality of components. The computing platform may identify human-engineered features of the URL. The computing platform may compute a vector representation of the URL to identify deep learned features of the URL. The computing platform may concatenate the human-engineered features of the URL to the deep learned features of the URL, resulting in a concatenated vector representation. By inputting the concatenated vector representation of the URL to a URL classifier, the computing platform may compute a phish classification score. In response to determining that the phish classification score exceeds a first phish classification threshold, the computing platform may cause a cybersecurity server to perform a first action.

Abstract: Systems and methods for fraud containment are provided in accordance with an embodiment of the invention. A fraud event may be detected. One or more devices that may be used in perpetrating the fraud event may be detected. Additionally one or more potential fraud victims, who may be grouped into victim circles may be detected. The threat level to the victims and/or victim circles may be assessed. In some instances, behavioral profiles may be utilized to make fraud assessments. Based on the threat level, recommendations for fraud containment responses may be provided.

Abstract: A computing system includes a processor, and a storage device holding instructions executable by the processor. The instructions are executable to receive a source code through an application programming interface (“API”) exposed to a graphical user interface (“GUI”). The GUI is configured to enable an author to define a proposed model with a pedagogical programming language, the proposed model including an input, one or more concept nodes, and an output. The GUI is further configured to enable the author to provide a program annotation indicating an execution behavior for the source code, to generate an assembly code from the source code with a compiler of an artificial intelligence (“AI”) engine configured to work with the GUI; and to build an executable, trained AI model including a neural-network layout having one or more layers derived from the assembly code.

Abstract: A processor-implemented method implementing a convolution neural network includes: determining a plurality of differential groups by grouping a plurality of raw windows of an input feature map into the plurality of differential groups; determining differential windows by performing, for each respective differential group of the differential groups, a differential operation between the raw windows of the respective differential group; determining a reference element of an output feature map corresponding to a reference raw window among the raw windows by performing a convolution operation between a kernel and the reference raw window; and determining remaining elements of the output feature map by performing a reference element summation operation based on the reference element and each of a plurality of convolution operation results determined by performing respective convolution operations between the kernel and each of the differential windows.

Abstract: A system including a processor may filter a population to identify a group having members, and deliver, over an electronic network, a survey to members of the group to determine value preferences and value gaps. The survey may be made according to a set of paired comparisons or other related techniques. The system may generate vectors according to the determined value preferences and value gaps for members of the group, which may have lengths equal to numbers of determined values for the members. The system may cluster members of the group into segments by calculating patterns of differences between generated vectors. Results of this process enable the creation of a predictive model to estimate segment membership, which can be used in the generation of a promotion for display corresponding to a clustered segment, and deliver the promotion to a client device associated for display.

Abstract: An information processing apparatus is communicably connected to a link rule storage unit. The link rule storage unit stores a link rule. The link rule represents a criterion as to whether or not to link a plurality of pieces of trajectory data representing a trajectory of an object into one. Further, the link rule is a criterion for linking the plurality of pieces of trajectory data estimated to represent the trajectory of the same object. The information processing apparatus obtains a plurality of pieces of trajectory data, and classifies the obtained plurality of pieces of trajectory data into groups for each object. The information processing apparatus updates the link rule stored in the link rule storage unit by using the plurality of pieces of trajectory data classified into the groups.

Abstract: Systems, methods, and computer-readable media are provided for determining treatments to apply to plants within control volumes having controlled agricultural environments. Each treatment comprises application of a set of setpoints, choosing a reproduction operation for the treatment, and selecting one or more previous sets from setpoints from one or more previously applied treatments, for use with the chosen reproduction operation.

Abstract: Described is a system for analyzing time series data. A sequence of symbols is generated from a set of time series input data related to a moving vehicle using automatic segmentation. A grammar is extracted from the sequence of symbols, and the grammar is a subset of a probabilistic context-free grammar (PCFG). Using the grammar, time series input data can be analyzed, and a prediction of the vehicle's movement can be made. Vehicle operations for an autonomous vehicle are determined using the prediction.

Abstract: A method is provided to create lead compound(s) by discovering a general chemical structure, moieties, formula(s) to explore suitable compositions by computer simulation and/or robotic biological or biochemical experiments at least partially based upon employing said lead compound(s) discover method, which includes steps for inputting at least one chemical formula and at least one byproduct formula, steps for creating a list of dipeptides that might dissociate the byproduct from the input formula by way of catalysis, steps for using these dipeptides to fingerprint a protein from its peptide sequence, and searching a protein database or use experimental methods to search for such proteins. A composition creating means is provide by way of computer simulation and/or robotic biological or biochemical experiments at least partially based upon employing, as lead compound(s), the final chemical structure, moieties, formula(s) generated and communicated the above method.

Abstract: A finance management platform and method are disclosed. A model element data repository encodes a plurality of model elements representing invoicing outcomes, each model element including a score value. A processor is configured to execute computer program code for providing a finance management platform for a plurality of clients, including receiving invoicing data from a client data repository remote from the finance management platform, translating the received invoicing data into a common format and store the invoicing data in the common format in an invoice data repository, accessing the model element data repository and determine one or more of the model elements applicable to the invoicing data, calculating an overall score for the invoicing data from the applicable model element's scores and triggering a communication to a payment processor and an update to the invoice data repository upon the overall score exceeding a predetermined threshold.

Abstract: An automated robot design pipeline facilitates the overall process of designing robots that perform various desired behaviors. The disclosed pipeline includes four stages. In the first stage, a generative engine samples a design space to generate a large number of robot designs. In the second stage, a metric engine generates behavioral metrics indicating a degree to which each robot design performs the desired behaviors. In the third stage, a mapping engine generates a behavior predictor that can predict the behavioral metrics for any given robot design. In the fourth stage, a design engine generates a graphical user interface (GUI) that guides the user in performing behavior-driven design of a robot. One advantage of the disclosed approach is that the user need not have specialized skills in either graphic design or programming to generate designs for robots that perform specific behaviors or express various emotions.

Abstract: Task/resources are randomly assigned a number of times and a score for each solution of the random assignment is calculated. Using machine learning and artificial intelligence, a subset of the solutions is selected. Assignment of task/resource within the subset may be randomly changed, e.g., a task/resource assignment between two entities, a task/resource within the selected subset may be replaced with another task/resource (without swapping), etc. The additional solutions form a super solution with the selected subset and the score associated with the additional solutions are calculated. The process of selection of assignments, random changes to the assignment and calculating the scores associated with the new solutions is repeated a number of times until a certain condition is met, e.g., a number of iterations, time out, improvement between two iterations is less than a certain threshold, etc. Once the certain condition is satisfied, a solution is selected.

Abstract: An example system can provide intelligent continuous learning by updating a machine learning model based on a new dataset. The system can utilize a transfer loss function that does not depend on old datasets used to train the existing model. The system can receive, on a graphical user interface (“GUI”), a selection of configuration criteria including threshold performance for automatic deployment. The new model can be created iteratively based on the configuration criteria. An evaluation of the new model over multiple iterations can be presented on the GUI. In an instance where the new model meets a deployment requirement selected on the GUI, a server can deploy the new model in place of the existing model.

Abstract: An all-encompassing system for planning and optimizing scheduled delivery appointments (or attended deliveries) through the utilization of a software tool having two major constituents: an asynchronously operating background advanced, genetic route optimization algorithm and an algorithm-informed communication user interface (whether through digital, or even analog, pathways, such as, without limitation, SMS, email, IM, and the like, oral communications, or any other possible means of communication). Such a system employs the optimization algorithm to determine efficient routes for delivery services in response to selections from contacted customers via the communication user interface. In this manner, the system allows for an initial base route generation from a set population of responding customers and negotiation through the communication user interface for a delivery time in relation to such a route.

Abstract: Data-dependent node-to-node knowledge sharing to increase the interpretability of the activation pattern of one or more nodes in a neural network, is implemented by a set of knowledge sharing links. Each link may comprise a knowledge providing node or other source P and a knowledge receiving node R. A knowledge sharing link can impose a node-specific regularization on the knowledge receiving node R to help guide the knowledge receiving node R to have an activation pattern that is more easily interpreted. The specification and training of the knowledge sharing links may be controlled by a cooperative human-AI learning supervisor system in which a human and an artificial intelligence system work cooperatively to improve the interpretability and performance of the client system.

Abstract: A control system for an operable system such as a flow control system or temperature control system. The system operates in a control loop to regularly update a model with respect at least one optimizable input variable based on the detected variables. The model provides prediction of use of the input variables in all possible operation points or paths of the system variables which achieve an output setpoint. In some example embodiments, the control loop is performed during initial setup and subsequent operation of the one or more operable elements in the operable system. The control system is self-learning in that at least some of the initial and subsequent parameters of the system are determined automatically during runtime.

Abstract: A user device that receives plural strings of data corresponding to physiological data, generated from a training model trained in an unsupervised manner using concurrently implemented loss functions, and using a distance metric, relative to a threshold, applied between the plural strings to provide an indication of authenticity.

Abstract: A method and system are provided for solving combinatorial optimization problems. A classical algorithm provides an approximate or “seed” solution which is then used by a quantum circuit to search its “neighborhood” for higher-quality feasible solutions. A continuous-time quantum walk (CTQW) is implemented on a weighted, undirected graph that connects the feasible solutions. An iterative optimizer tunes the quantum circuit parameters to maximize the probability of obtaining high-quality solutions from the final state. The ansatz circuit design ensures that only feasible solutions are obtained from the measurement. The disclosed method solves constrained problems without modifying their cost functions, confines the evolution of the quantum state to the feasible subspace, and does not rely on efficient indexing of the feasible solutions as some previous methods require.

Abstract: One embodiment provides a method, including: identifying, utilizing one or more sensors operatively coupled to the smart device, a user operating a smart device; inferring, based upon the identifying the user, a knowledge level of the user with respect to the smart device; determining, based in part on the knowledge level of the user, at least one input within a plurality of inputs provided by the user is anomalous with respect to a pattern of inputs utilized to execute a command; transforming, based upon the anomalous at least one input and the pattern, the plurality of inputs to a plurality of inputs matching the pattern of inputs; and executing an action based upon the transformed plurality of inputs.

Abstract: Systems and methods are provided for operating to an initial optimized baseline solution to a multi-objective problem. As the initial optimized baseline solution is determined, some regions, such as local or global maxima, minima, and/or saddle points in the objective space may be mapped. The mapping may be performed by storing mesh chromosomes corresponding to some of the features (e.g., extrema, saddle points, etc.) in the objective space along with the location of those chromosomes within the objective space (e.g., objective values corresponding to each of the objectives). The mesh chromosome may be used in subsequent re-optimization problems, such as with reformulation. Although in a re-optimization the objectives, decision variables, and or objective/constraint models may change, the mesh chromosomes may still provide information and direction for more quickly and/or with reduced resources converge on a re-optimized solution.

Abstract: A central hub and database for a smart home environment enable the learning of states associated with items within the smart home and the training one or more machine-learned models associated with the items. After training the machine-learned models, the central hub can modify a state of an item based on the machine-learned model associated with the item. For instance, a window can be opened or shut, a light can be dimmed or turned off, and a door can be locked. Each state of the object can be associated with a set of conditions that, when satisfied, cause the central hub to change the state of the object using the corresponding machine-learned model, for instance without receiving an explicit input from a user.

Abstract: Roughly described, a problem solving platform distributes the solving of the problem over a evolvable individuals, each of which also evolves its own pool of actors. The actors have the ability to contribute collaboratively to a solution at the level of the individual, instead of each actor being a candidate for the full solution. Populations evolve both at the level of the individual and at the level of actors within an individual. In an embodiment, an individual defines parameters according to which its population of actors can evolve. The individual is fixed prior to deployment to a production environment, but its actors can continue to evolve and adapt while operating in the production environment. Thus a goal of the evolutionary process at the level of individuals is to find populations of actors that can sustain themselves and survive, solving a dynamic problem for a given domain as a consequence.

Abstract: Systems and methods for forecasting future values of data streams are disclosed. One example method may include receiving information characterizing each of a plurality of forecasting models, retrieving historical data for each of a plurality of data streams, determining one or more constraints, dynamically selecting one of the plurality of forecasting models for each of the data streams based on accuracy metrics for the forecasting models, estimating cost metrics associated with each forecasting model, dynamically selecting the forecasting model based at least in part on the accuracy metrics, the cost metrics, and the determined constraints, and forecasting a first subsequent value of each data stream using the corresponding selected forecasting model.

Abstract: A system for localizing a transmitting wireless device within a known volume, the system including antennae deployed in respective locations at least some within the known volume, each of the antennae operative to receive and output a signal from the transmitting device; analog-to-digital converter/s; and a processor receiving digital sampled received signals and computing real time output parameter/s as function of: sampled received signals S, received from the transmitting wireless device at antenna i; and of sampled signals, received from the transmitting wireless device at antenna j and sampled, simultaneously with reception at antenna I and sampling of sampled received signals S, function being independent of power level at which transmitting device is transmitting, and to estimate transmitting wireless device's location within volume by comparing the parameter/s to reference output parameters respectively having known correspondence to known possible locations/s within volume, for at least two of the an

Abstract: Techniques facilitating estimating and visualizing entity to agent collaboration to facilitate automated plan generation are provided. In one example, a computer-implemented method comprises generating, by a device operatively coupled to a processor, a plan based on receiving first input data associated with an instance model. The computer-implemented method also comprises generating, by the device, a revised plan based on receiving second input data, associated with a revised instance model, from an entity. Furthermore, the computer-implemented method comprises, tracking, by the device, a contribution of the entity as a function of a modification from the instance model to the revised instance model.

Abstract: Aspects of the disclosure relate to detecting and identifying malicious sites using machine learning. A computing platform may receive a uniform resource locator (URL). The computing platform may parse and/or tokenize the URL to reduce the URL into a plurality of components. The computing platform may identify human-engineered features of the URL. The computing platform may compute a vector representation of the URL to identify deep learned features of the URL. The computing platform may concatenate the human-engineered features of the URL to the deep learned features of the URL, resulting in a concatenated vector representation. By inputting the concatenated vector representation of the URL to a URL classifier, the computing platform may compute a phish classification score. In response to determining that the phish classification score exceeds a first phish classification threshold, the computing platform may cause a cybersecurity server to perform a first action.

Abstract: A system for a generating an alimentary element prediction machine-learning model, the system comprising a computing device configured to provide, to a user, a plurality of compatible alimentary elements as a function of user biochemistry, receive training data relating a plurality of temporally preceding alimentary elements as a function of the plurality of compatible alimentary elements presented to a user, train, using a machine-learning process, a computer model as a function of the user-selection training data to predict user-selectable alimentary elements, generate an alimentary profile as a function of the computer model, receive a user input for an alimentary element, and present, as a function of the user input, the alimentary element as a function of the alimentary profile.

Abstract: A visible light sensor may be configured to sense environmental characteristics of a space using an image of the space. The visible light sensor may be controlled in one or more modes, including a daylight glare sensor mode, a daylighting sensor mode, a color sensor mode, and/or an occupancy/vacancy sensor mode. In the daylight glare sensor mode, the visible light sensor may be configured to decrease or eliminate glare within a space. In the daylighting sensor mode and the color sensor mode, the visible light sensor may be configured to provide a preferred amount of light and color temperature, respectively, within the space. In the occupancy/vacancy sensor mode, the visible light sensor may be configured to detect an occupancy/vacancy condition within the space and adjust one or more control devices according to the occupation or vacancy of the space. The visible light sensor may be configured to protect the privacy of users within the space via software, a removable module, and/or a special sensor.

Abstract: A computer-implemented method that, in an embodiment, includes receiving a travel request from a traveler that includes a pickup and destination location and a number of travelers and analyzing the travel request and calculating routes for vehicles that are partially based on the pickup and destination location, the number of travelers, destination locations of travelers located in the vehicles, traffic conditions, minimizing a waiting time for the traveler, minimizing a travel time for the traveler, minimizing a travel time for the travelers located in the vehicles, and optimizing an aggregation of travelers on each of the plurality vehicles. In an embodiment, the plurality of vehicles includes 5 vehicles or more and based on the calculating, the method includes supplying a selected route to a selected vehicle and supplying to the traveler, identification information related to the selected vehicle thereby resulting in transportation of the traveler to the destination location.

Abstract: A system and method for adapting an errant automated decisioning workflow includes reconfiguring digital abuse or digital fraud logic parameters associated with automated decisioning routes of an automated decisioning workflow in response to identifying an anomalous drift or an anomalous shift in efficacy metrics of the automated decisioning workflow, wherein the automated decisioning workflow includes a plurality of distinct automated decisioning routes that, when applied in a digital threat evaluation of data associated with a target digital event, automatically compute a decision for disposing the target digital event based on a probability digital fraud; simulating, by computers, a performance of the automated decisioning routes in a reconfigured state based on inputs of historical digital event data; calculating simulation metrics based on simulation output data of the simulation; and promoting to an in-production state the automated decisioning workflow having the automated decisioning routes in the rec

Abstract: A method for providing automated predictive analysis for a case management system is disclosed and includes providing, by a server, a case management system comprising configuration information defining a plurality of case attributes and performance criteria comprising at least one performance criterion, and comprising case management data associated with the plurality of case attributes and a plurality of cases, and automatically determining a projected outcome associated with a performance criterion included in the configuration information. The method also includes identifying a correlation between a case attribute and the projected outcome based on case management data associated with the case attribute and/or case management data associated with the performance criterion, generating a prediction model based on the correlation and providing a prediction relating to the projected outcome based on the correlation.

Abstract: Methods and systems for improving vehicular safety by utilizing uplift modeling techniques to improve a driving tip treatment generation model are provided. According to embodiments, a tip server can analyze telematics data associated with operation of one or more vehicles to determine that a driving tip should be provided to a driver of a vehicle. The tip server then utilize a treatment generation model to determine a treatment for how to provide the driving tip in a manner optimized for the particular driver. The tip server can analyze additional telematics data to determine an effectiveness of the driving tip and to update the treatment generation model in accordance with uplift modeling techniques.

Abstract: The present disclosure relates to a system, a method, and a product for optimizing hyper-parameters for generation and execution of a machine-learning model under constraints. The system includes a memory storing instructions and a processor in communication with the memory. When executed by the processor, the instructions cause the processor to obtain input data and an initial hyper-parameter set; for an iteration, to build a machine learning model based on the hyper-parameter set, evaluate the machine learning model based on the target data to obtain a performance metrics set, and determine whether the performance metrics set satisfies the stopping criteria set. If yes, the instructions cause the processor to perform an exploitation process to obtain an optimal hyper-parameter set, and exit the iteration; if no, perform an exploration process to obtain a next hyper-parameter set, and perform a next iteration with using the next hyper-parameter set as the hyper-parameter set.

Abstract: A machine and its modules assist in steganography for an animal. A steganography module applies behavioral sequencing to create a cover message and a hidden message to covertly pass information from one animal to another animal, with the information embedded in an individual's brain. A visual module references the steganography module to cause a sequence of visual images on a display screen to guide a motor sequence of an individual as the cover message as well as detect and communicate a timing of the individual's motor sequence, relative in timing, to visual images in the sequence of visual images being displayed on the display screen, in order to train in the cover message and hidden message. The hidden message is then extracted at a destination from a sensor monitoring the individual's sequence of motor actions.

Abstract: Methods, systems and apparatus for simulating quantum circuits including multiple quantum logic gates. In one aspect, a method includes the actions of representing the multiple quantum logic gates as functions of one or more classical Boolean variables that define a undirected graphical model with each classical Boolean variable representing a vertex in the model and each function of respective classical Boolean variables representing a clique between vertices corresponding to the respective classical Boolean variables; representing the probability of obtaining a particular output bit string from the quantum circuit as a first sum of products of the functions; and calculating the probability of obtaining the particular output bit string from the quantum circuit by directly evaluating the sum of products of the functions. The calculated partition function is used to (i) calibrate, (ii) validate, or (iii) benchmark quantum computing hardware implementing a quantum circuit.

Abstract: Various systems and methods of aggregating and analyzing geographically indexed data are disclosed. The system can include a server database hosting an application that a client computer may access via a web browser according to a SaaS architecture. The server database can store a variety of geographically indexed data, which may include economic data, demographic data, social data, and various other data types. The server database can be programmed to cause the client to display a map for receiving a selected geographic area defined on the map and then retrieve selected data corresponding to user-selected criteria for the selected geographic area. The server database can then transmit the selected data to the client for display on the map.

Abstract: Method for improving user authentication efficiency performed by a communication device belonging to an authentication system. The communication device includes a local machine learning engine having a set of N artificial neural network ANN1,i adapted to process N different types of input signals.

Abstract: System and method for generating case-based data including receiving, input data describing an event of interest; if the input data is not in a format of a property graph then transforming the input data into a first property graph describing the event of interest and representing a first network, wherein the first property graph includes a plurality of network elements and properties of at least some of the plurality of network elements, wherein the network elements include entities and links describing relationships between the entities; changing a network element in the first property graph to create a second property graph of a new network; and using the second property graph as the case-based data. New properties may be generated for the entities and links.

Abstract: Embodiments of the present disclosure relate to a method, a device and a computer-readable storage medium for data processing. The method for data processing comprises: obtaining a set of observation samples regarding a plurality of factors, one of the set of observation samples comprising respective observed values of the plurality of factors. The method further comprises: estimating, for each of the plurality of factors and based on the set of observation samples, a distribution that differences between observed values of the factor and estimated values of the factor follow. The method further comprises determining, based at least on the estimated distribution, a causal structure representing a causal relationship among the plurality of factors. Embodiments of the present disclosure further provide a device and a computer-readable storage medium for implementing the above method.

Abstract: A composite novelty method approach to deceptive problems where a secondary objective is available to diversify the search is described. In such cases, composite objectives focus the search on the most useful tradeoffs and allow escaping deceptive areas. Novelty-based selection increases exploration in the focus area, leading to better solutions, faster and more consistently and it can be combined with other fitness-based methods.

Abstract: An embodiment of the present invention is directed to allowing permissioned legal entity managers to define rules for a legal entity to derive an association of an application with a legal entity. The application manager may then be presented with an option to accept suggested associations or perform an override. Once a legal entity is ready, applications may be published to a system of record. Due to the vast amount of applications and entities available, the solution provides automation while still allowing overrides. The results of the classification may then be stored in a system of record and data quality suggestions sent to the source systems to clean source system data. An embodiment of the present invention may also associate other attributes to applications and deployments.

Abstract: Described herein is a process which facilitates segmented security between domain-specific data sets being evaluated as part of a candidate evaluation service and third-party evolution services, wherein the data sets are not transmitted to the evolution service which is evolving candidates for evaluation. This enables customers with secure data sets to use candidate evolution services securely by obtaining a population of potentially optimal candidate models to evaluate and then optimizing on those data sets in their own secure fashion.

Abstract: An optimization apparatus includes an output data acquirer to acquire output data expressing a result of experiment or simulation based input parameters, input/output data storage to store the input parameters and the output data corresponding to the input parameters, as a pair, an evaluation value calculator to calculate evaluation values of the output data, an input parameter converter to generate conversion parameters of a dimension number changed from the dimension number of the input parameters, a next-input parameter decider to decide next input parameters based on pairs of the conversion parameters and the evaluation values corresponding to the conversion parameters, and a repetition determiner to repeat operations of the output data acquirer, the input/output data storage, the evaluation value calculator, the input parameter converter, and the next-input parameter decider, until satisfying a predetermined condition.

Abstract: An information processing device according to one embodiment includes a processor. The processor executes an acquisition step of acquiring a plurality of partial layers as a gene pool, the partial layers being candidates for elements of a deep learning model, a generation step of generating a new partial layer from the gene pool by using genetic programming, an evaluation step of evaluating each partial layer by incorporating, into a template of the deep learning model, each partial layer in the gene pool to which the new partial layer is added, and determining a plurality of partial layers to remain in the gene pool, and an output step of outputting the deep learning model into which the partial layer with a highest evaluation value is incorporated.

Abstract: One or more processors access a plurality of messages from a discussion community and determine concept topics and metadata of the messages. A graph of the concept topics and messages is generated with messages and concept topics as nodes and connection strength between nodes as edges. Distances between nodes correspond to computed weights of edges. A low boundary distance threshold is calculated based on the aggregate distances between existing messages and concept topics of the graph. A distance is computed between new concept topics of a received message and existing messages and concept topics. A distance between the new concept topics of the received message and existing nodes is determined and compared to the boundary distance threshold of the graph, and responsive to determined distance falling below the boundary threshold, an action is performed to anchor the new concept topics of the received message in the discussion community.

Abstract: A method for ranking genealogical records includes using a machine learning model to rank multiple searched records based on relevancy. The relevancy may be determined by identifying features included in a record and scaling each feature by a corresponding weight factor. In addition, a method for training a machine learning model and increasing the convergence speed of the training is described. To train the model, a machine learning process is used to optimize a ranking performance metric. A set of weights corresponding to multiple features are used to rank multiple past search records in a training set. An initial set of the weights are set by the expectation values of the weights. The weights are incrementally changed to optimize the ranking performance metric. The step size of the increment is determined based on the sensitivity of the ranking performance metric relative to the step size.

Abstract: A method of classifying design criteria includes receiving design criteria for a product part. The criteria comprise one or both of performance and manufacturing criteria. The design criteria are sorted into different classes of one or both of one or more objective functions and one or more constraints based on when they can be satisfied or optimized. Constraint violations are determined. A design workflow is produced to generate one or more designs of a part to comply with one or more of satisfying constraints and optimizing objective functions.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for adversarial training of a neural network. One of the methods includes obtaining a plurality of training inputs; and training the neural network on each of the training inputs, comprising, for each of the training inputs: processing the training input using the neural network to determine a neural network output for the training input; applying a perturbation to the training input to generate an adversarial perturbation of the training input; processing the adversarial perturbation of the training input using the neural network to determine a neural network output for the adversarial perturbation; and adjusting the current values of the parameters of the neural network by performing an iteration of a neural network training procedure to optimize an adversarial objective function.

Abstract: There is disclosed in one example a computing apparatus, including: a hardware platform including a processor and a memory; a network interface; a user-space application including instructions to interact with a web site via a uniform resource locator (URL); and a security agent including instructions to: intercept an interaction of the user-space application with the web site; determine that the intercepted interaction is to send sensitive information to the web site; suspend the interaction; and assign a reputation to the URL.