A neural network system is proposed. The neural network can be trained by model-based reinforcement learning to select actions to be performed by an agent interacting with an environment, to perform a task in an attempt to achieve a specified result. The system may comprise at least one imagination core which receives a current observation characterizing a current state of the environment, and optionally historical observations, and which includes a model of the environment. The imagination core may be configured to output trajectory data in response to the current observation, and/or historical observations. The trajectory data comprising a sequence of future features of the environment imagined by the imagination core. The system may also include a rollout encoder to encode the features, and an output stage to receive data derived from the rollout embedding and to output action policy data for identifying an action based on the current observation.

Methods, systems, and devices for determining a subset of user devices from among a complete set of user devices based on a set of received information, i.e., attributes associated with a photograph or user device that transmitted the photograph and attributes, where the disposition of the information may be used to determine the subset and then perform facial recognition on the subset of user associated photographs in order to accurately identify each user or users present in the photograph.

Methods, systems, and devices for determining a subset of user devices from among a complete set of user devices based on a set of received information, i.e., attributes associated with a photograph or user device that transmitted the photograph and attributes, where the disposition of the information may be used to determine the subset and then perform facial recognition on the subset of user associated photographs in order to accurately identify each user or users present in the photograph.

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 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 method of identifying and recognizing characters using a dual-stage neural network pipeline, the method including: receiving, by a computing device, image data; providing the image data to a first convolutional layer of a convolutional neural network (CNN); applying, using the CNN, pattern recognition to the image data to identify a region of the image data containing text; providing sub-image data comprising the identified region of the image data to a convolutional recurrent neural network (CRNN); and recognizing, using the CRNN, the characters within the sub-image data.

In various embodiments, a first plurality of digital images captured by a first camera (256, 456, 1156) of a first area may be categorized (1202-1210) into multiple predetermined categories based on visual attribute(s) of the first plurality of digital images. A second plurality of digital images captured by a second camera (276, 376, 476, 1176) of a second area may be categorized (1302-1310) into the same predetermined categories based on visual attribute(s) of the second plurality of digital images. After the second camera acquires (1402) a subsequent digital image depicting an unknown subject in the second area, the subsequent digital image may be categorized (1404-1406) into a given one of the predetermined categories based on its visual attribute(s), and then adjusted (1408) based on a relationship between the first plurality of digital images categorized into the given category and the second plurality of digital images categorized into the given category.

A method for detecting comparison persons 7 to a search person 4, wherein a plurality of classification persons 3 is classified by extracting values W1,W2,W3 for classification features K1,K2,K3 from classification images 2 of the classification persons 3, the classification being ambiguous in such a way that the classification does not enable a unique identification of any of the classification persons 3, wherein during a search for a search person 4 using a search image 5 by a comparison of values of search features from the search image 5 with values W1,W2,W3 of classification features K1,K2,K3, at least two classification persons 3 are output as comparison persons 7.

Methods, systems, and computer-readable storage media for receiving, by an automated review system, a legal document as a computer-readable file, and determining, by the automated review system, that the legal document is of a first type, and in response: converting the legal document to a set of images, extracting text data from one or more images in the set of images, the text data including sub-sets of text data, each sub-set of text data representing text in a respective clause of a set of clauses of the legal document, for each sub-set of text data receiving a prediction from a machine learning (ML) model in a set of ML models, the ML model being specific to a clause in the set of clauses, and outputting a set of predictions and respective prediction values for display in a user interface (UI).

Methods, systems, and devices for determining a subset of user devices from among a complete set of user devices based on a set of received information, i.e., attributes associated with a photograph or user device that transmitted the photograph and attributes, where the disposition of the information may be used to determine the subset and then perform facial recognition on the subset of user associated photographs in order to accurately identify each user or users present in the photograph.