Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling an agent. One of the methods includes receiving a current observation characterizing a current state of the environment as of the time step; generating an embedding of the current observation; processing scene memory data comprising embeddings of prior observations received at prior time steps using an encoder neural network, wherein the encoder neural network is configured to apply an encoder self-attention mechanism to the scene memory data to generate an encoded representation of the scene memory data; processing the encoded representation of the scene memory data and the embedding of the current observation using a decoder neural network to generate an action selection output; and causing the agent to perform the selected action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling an agent. One of the methods includes receiving a current observation characterizing a current state of the environment as of the time step; generating an embedding of the current observation; processing scene memory data comprising embeddings of prior observations received at prior time steps using an encoder neural network, wherein the encoder neural network is configured to apply an encoder self-attention mechanism to the scene memory data to generate an encoded representation of the scene memory data; processing the encoded representation of the scene memory data and the embedding of the current observation using a decoder neural network to generate an action selection output; and causing the agent to perform the selected action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling an agent. One of the methods includes receiving a current observation characterizing a current state of the environment as of the time step; generating an embedding of the current observation; processing scene memory data comprising embeddings of prior observations received at prior time steps using an encoder neural network, wherein the encoder neural network is configured to apply an encoder self-attention mechanism to the scene memory data to generate an encoded representation of the scene memory data; processing the encoded representation of the scene memory data and the embedding of the current observation using a decoder neural network to generate an action selection output; and causing the agent to perform the selected action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling an agent. One of the methods includes receiving a current observation characterizing a current state of the environment as of the time step; generating an embedding of the current observation; processing scene memory data comprising embeddings of prior observations received at prior time steps using an encoder neural network, wherein the encoder neural network is configured to apply an encoder self-attention mechanism to the scene memory data to generate an encoded representation of the scene memory data; processing the encoded representation of the scene memory data and the embedding of the current observation using a decoder neural network to generate an action selection output; and causing the agent to perform the selected action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: An example method is disclosed that includes identifying a training set of images, wherein each image in the training set has an identified bounding box that comprises an object class and an object location for an object in the image. The method also includes segmenting each image of the training set, wherein segments comprise sets of pixels that share visual characteristics, and wherein each segment is associated with an object class. The method further includes clustering the segments that are associated with the same object class, and generating a data structure based on the clustering, wherein entries in the data structure comprise visual characteristics for prototypical segments of objects having the object class and further comprise one or more potential bounding boxes for the objects, wherein the data structure is usable to predict bounding boxes of additional images that include an object having the object class.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: Digital graphic novel content is received and features of the graphic novel content are identified. At least one of the identified features includes text. Contextual information corresponding to the feature or features that include text is generated based on the identified features. The contextual information is used to aid translation of the text included in the feature or features that include text.

Abstract: A linear function describing a framework for identifying an object of class k in an image sample x may be described by: wk*x+bk, where bk is the bias term. The higher the value obtained for a particular classifier, the better the match or strength of identity. A method is disclosed for classifier and/or content padding to convert dot-products to distances, applying a hashing and/or nearest neighbor technique on the resulting padded vectors, and preprocessing that may improve the hash entropy. A vector for an image, an audio, and/or a video may be received. One or more classifier vectors may be obtained. A padded image, video, and/or audio vector and classifier vector may be generated. A dot product may be approximated and a hashing and/or nearest neighbor technique may be performed on the approximated dot product to identify at least one class (or object) present in the image, video, and/or audio.

Abstract: Digital graphic novel content is received and features of the graphic novel content are identified. At least one of the identified features includes text. Contextual information corresponding to the feature or features that include text is generated based on the identified features. The contextual information is used to aid translation of the text included in the feature or features that include text.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: Digital graphic novel content is received and a machine-learning model applied to predict features of the digital graphic novel content. The predicted features include locations of a plurality of panels and a reading order of the plurality of panels. A packaged digital graphic novel is created that includes the digital graphic novel content and presentation metadata. The presentation metadata indicates a manner in which the digital graphic novel content should be presented based on the locations and reading order of the plurality of panels. The packaged digital graphic novel is provided to a reading device to be presented in accordance with the manner indicated in the presentation metadata.

Abstract: Digital graphic novel content is received and features of the graphic novel content are identified. At least one of the identified features includes text. Contextual information corresponding to the feature or features that include text is generated based on the identified features. The contextual information is used to aid translation of the text included in the feature or features that include text.

Abstract: Systems and techniques are provided for a ranking approach to train deep neural nets for multilabel image annotation. Label scores may be received for labels determined by a neural network for training examples. Each label may be a positive label or a negative label for the training example. An error of the neural network may be determined based on a comparison, for each of the training examples, of the label scores for positive labels and negative labels for the training example and a semantic distance between each positive label and each negative label for the training example. Updated weights may be determined for the neural network based on a gradient of the determined error of the neural network. The updated weights may be applied to the neural network to train the neural network.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating descriptions of input images. One of the methods includes obtaining an input image; processing the input image using a first neural network to generate an alternative representation for the input image; and processing the alternative representation for the input image using a second neural network to generate a sequence of a plurality of words in a target natural language that describes the input image.

Abstract: A linear function describing a framework for identifying an object of class k in an image sample x may be described by: wk*x+bk, where bk is the bias term. The higher the value obtained for a particular classifier, the better the match or strength of identity. A method is disclosed for classifier and/or content padding to convert dot-products to distances, applying a hashing and/or nearest neighbor technique on the resulting padded vectors, and preprocessing that may improve the hash entropy. A vector for an image, an audio, and/or a video may be received. One or more classifier vectors may be obtained. A padded image, video, and/or audio vector and classifier vector may be generated. A dot product may be approximated and a hashing and/or nearest neighbor technique may be performed on the approximated dot product to identify at least one class (or object) present in the image, video, and/or audio.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for detecting objects in images. One of the methods includes receiving an input image. A full object mask is generated by providing the input image to a first deep neural network object detector that produces a full object mask for an object of a particular object type depicted in the input image. A partial object mask is generated by providing the input image to a second deep neural network object detector that produces a partial object mask for a portion of the object of the particular object type depicted in the input image. A bounding box is determined for the object in the image using the full object mask and the partial object mask.