The technology described herein is directed to a media indexer framework including a character recognition engine that automatically detects and groups instances (or occurrences) of characters in a multi-frame animated media file. More specifically, the character recognition engine automatically detects and groups the instances (or occurrences) of the characters in the multi-frame animated media file such that each group contains images associated with a single character. The character groups are then labeled and used to train an image classification model. Once trained, the image classification model can be applied to subsequent multi-frame animated media files to automatically classifying the animated characters included therein.

The present disclosure relates to a method for generating control signals to assist occupants in a vehicle, wherein a context of the vehicle is determined, a rule of a rule-based data system is selected depending on the determined context, wherein the rule-based data system comprises a plurality of rules, wherein each rule has a condition part and a result part, wherein the condition part comprises conditions for the context of the vehicle, a confidence value associated with the selected rule is determined, wherein the confidence value indicates the probability with which the result of the rule corresponds with the preference of the user, a result of the selected rule is generated, a control signal is generated and output depending on the generated rule result, wherein the control signal automates a vehicle function with a degree of automation, wherein the degree of automation depends on the confidence value of the selected rule. The disclosure likewise relates to a device for executing this method.

This specification describes a computer-implemented method for testing the performance of a video game server. The method comprises initializing a recurrent neural network. The recurrent neural network is trained based on requests sent from one or more client devices to the video game server. The initializing comprises inputting a start token into the recurrent neural network. An output distribution for a first-time step is generated, as an output of the recurrent neural network. The output distribution comprises a probability of generating each of a set of one or more requests to the video game server, in addition to a probability of generating a stop token. A first request from the set of one or more requests is selected based on the output distribution. The method comprises for one or more further time steps, until a stop token has been selected from output of the recurrent neural network: inputting, into the recurrent neural network, a request selected in the previous time step; generating, as an output of the recurrent neural network, an output distribution for the time step; and selecting, based on the output distribution, a request. A generated sequence of requests is stored. The generated sequence of requests comprises one or more of the requests selected at each respect time step. The generated sequence of requests is inputted into a test generator. A performance test for testing the performance of the video game server is generated by the test generator.

A method for training a control strategy with the aid of reinforcement learning. The method includes carrying out passes, in each pass, an action that is to be carried out being selected for each state of a sequence of states of an agent, for at least some of the states the particular action being selected by specifying a planning horizon that predefines a number of states, ascertaining multiple sequences of states, reachable from the particular state, using the predefined number of states, by applying an answer set programming solver to an answer set programming program which models the relationship between actions and the successor states that are reached by the actions, selecting the sequence that delivers the maximum return, and selecting an action as the action for the particular state via which the first state of the selected sequence may be reached, starting from the particular state.

A network authentication system can be configured for sampling a plurality of signal samples from a device on a network, providing the plurality of signal samples to a first machine-learned model that is configured to determine a device fingerprint based at least in part on the plurality of signal samples, and providing the device fingerprint to a second machine-learned model that is configured to classify the device based at least in part on the device fingerprint.

A computing system may automatically classify applications that are used via a communication network. Application classification may include identifying a signature or group of signatures that belongs to an application or service associated with data flow through a network. The computer system of the network may collect data regarding the application from a mobile device, from the network, and/or from a digital distribution service accessible via the network. The system may combine such data together to identify and classify the application.

Methods, computer program products and/or systems are provided that perform the following operations: obtaining a first graph comprising first nodes representing first entities and first edges representing relationships between first entities, the first nodes being associated with first entity attributes descriptive of the first entities represented by the first nodes, the first edges being associated with first edge attributes descriptive of the relationships represented by the first edges; determining a first subgraph for a certain node of the first nodes of the first graph, the first subgraph including the certain node and at least one neighboring node of the certain node; and determining a data quality issue regarding the certain node based, at least in part, on applying one or more applicable rules of a set of data quality rules to first entity attribute values and first edge attribute values of the first subgraph.

Storage management is performed. For example, a computing device may determine that a fault belongs to one of a plurality of predefined fault categories based on description information of the fault of a storage system. Then, the computing device may determine at least one fault cause associated with the fault category at a first level of a hierarchical structure of predetermined fault causes. Further, the computing device may determine a first fault cause that causes the fault among the at least one fault cause. After that, the computing device may determine a target fault cause at the deepest level that causes the fault based on the first fault cause. As a result, the root cause of a fault of a storage system may be accurately and efficiently determined, thereby providing the possibility of fundamentally eliminating the fault.

A method including training predictor machine learning models (MLMs) using a first data set. The trained predictor MLMs are trained to predict classifications of data items in the first data set. The method also includes training confidence MLMs using second classifications, output by the trained predictor MLMs. The method also includes generating an aggregated ranked list of classes based on third classifications output by the trained predictor MLMs and second confidences output by the trained confidence MLMs. The method also includes training an ensemble confidence MLM using the aggregated ranked list of classes to generate a trained ensemble confidence MLM. The trained ensemble confidence MLM is trained to predict a corresponding selected classification for each corresponding data item in a training data set containing second data items similar to the first data items.

A novel distributed method for machine learning is described, where the algorithm operates on a plurality of data silos, such that the privacy of the data in each silo is maintained. In some embodiments, the attributes of the data and the features themselves are kept private within the data silos. The method includes a distributed learning algorithm whereby a plurality of data spaces are co-populated with artificial, evenly distributed data, and then the data spaces are carved into smaller portions whereupon the number of real and artificial data points are compared. Through an iterative process, clusters having less than evenly distributed real data are discarded. A plurality of final quality control measurements are used to merge clusters that are too similar to be meaningful. These distributed quality control measures are then combined from each of the data silos to derive an overall quality control metric.