Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing images or features of images using an image classification system that includes a batch normalization layer. One of the systems includes a convolutional neural network configured to receive an input comprising an image or image features of the image and to generate a network output that includes respective scores for each object category in a set of object categories, the score for each object category representing a likelihood that that the image contains an image of an object belonging to the category, and the convolutional neural network comprising: a plurality of neural network layers, the plurality of neural network layers comprising a first convolutional neural network layer and a second neural network layer; and a batch normalization layer between the first convolutional neural network layer and the second neural network layer.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing images or features of images using an image classification system that includes a batch normalization layer. One of the systems includes a convolutional neural network configured to receive an input comprising an image or image features of the image and to generate a network output that includes respective scores for each object category in a set of object categories, the score for each object category representing a likelihood that that the image contains an image of an object belonging to the category, and the convolutional neural network comprising: a plurality of neural network layers, the plurality of neural network layers comprising a first convolutional neural network layer and a second neural network layer; and a batch normalization layer between the first convolutional neural network layer and the second neural network layer.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for implementing long-short term memory cells with saturating gating functions. One of the systems includes a first Long Short-Term Memory (LSTM) cell, wherein the first LSTM cell is configured to, for each of the plurality of time steps, generate a new cell state and a new cell output by applying a plurality of gates to a current cell input, a current cell state, and a current cell output, each of the plurality of gates being configured to, for each of the plurality of time steps: receive a gate input vector, generate a respective intermediate gate output vector from the gate input, and apply a respective gating function to each component of the respective intermediate gate output vector, wherein the respective gating function for at least one of the plurality of gates is a saturating gating function.

Abstract: A computer-implemented method, computer program product, and computing system is provided for interacting with images having similar content. In an embodiment, a method may include identifying a plurality of photographs as including a common characteristic. The method may also include generating a flipbook media item including the plurality of photographs. The method may further include associating one or more interactive control features with the flipbook media item.

Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.

Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for implementing a neural network. In one aspect, the neural network includes a batch renormalization layer between a first neural network layer and a second neural network layer. The first neural network layer generates first layer outputs having multiple components. The batch renormalization layer is configured to, during training of the neural network on a current batch of training examples, obtain respective current moving normalization statistics for each of the multiple components and determine respective affine transform parameters for each of the multiple components from the current moving normalization statistics. The batch renormalization layer receives a respective first layer output for each training example in the current batch and applies the affine transform to each component of a normalized layer output to generate a renormalized layer output for the training example.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer.

Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

Abstract: Methods, systems, and media for summarizing a video with video thumbnails are provided.

Abstract: Methods, systems, and media for summarizing a video with video thumbnails are provided.

Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

Abstract: Methods and systems are disclosed for estimating a user's ability to share content that is of interest to recipients, and of informing a recipient of this ability when the user shares content with the recipient. In one embodiment, a computer system receives an indication that a first user wishes to share a content item (e.g., a video clip, a photo, an audio clip, a webpage, etc.) with a second user. In response, the computer system obtains data pertaining to a prior history of interaction by the second user with content that the first user has previously shared with the second user; determines, based on the obtained data, an estimate of an ability of the first user to predict an interest in the content item by the second user; and provides the estimate to the second user.

Abstract: Methods, systems, and media for summarizing a video with video thumbnails are provided.

Abstract: A method includes dividing a video uploaded to a user's client device into scenes that include one or more frames. The method also includes generating a digital summary for each scene based on content associated with a respective portion of the video spanned by the scene. The method also includes identifying a matching portion of the uploaded video containing third-party content base on a match between the digital summary associated with the matching portion and the digital summary associated with the third-party content. The method also includes identifying an original portion of the video containing the original content and a usage policy associated with a content owner of the third-party content. The method also includes generating a degraded video based on the usage policy, by applying a quality reduction to the matching portion.

Abstract: An image processing system automatically segments and labels an image using a hierarchical classification model. A global classification model determines initial labels for an image based on features of the image. A label-based descriptor is generated based on the initial labels. A local classification model is then selected from a plurality of learned local classification model based on the label-based descriptor. The local classification model is applied to the features of the input image to determined refined labels. The refined labels are stored in association with the input image.

Abstract: Methods, systems, and media for summarizing a video with video thumbnails are provided.

Abstract: A method includes dividing a video uploaded to a user's client device into scenes that include one or more frames. The method also includes generating a digital summary for each scene based on content associated with a respective portion of the video spanned by the scene. The method also includes identifying a matching portion of the uploaded video containing third-party content base on a match between the digital summary associated with the matching portion and the digital summary associated with the third-party content. The method also includes identifying an original portion of the video containing the original content and a usage policy associated with a content owner of the third-party content. The method also includes generating a degraded video based on the usage policy, by applying a quality reduction to the matching portion.