Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. One of the methods includes processing each training input using the neural network and in accordance with the current values of the network parameters to generate a network output for the training input; computing a respective loss for each of the training inputs by evaluating a loss function; identifying, from a plurality of possible perturbations, a maximally non-linear perturbation; and determining an update to the current values of the parameters of the neural network by performing an iteration of a neural network training procedure to decrease the respective losses for the training inputs and to decrease the non-linearity of the loss function for the identified maximally non-linear perturbation.

In one embodiment, a set of feature vectors can be derived from any biometric data, and then using a deep neural network (“DNN”) on those one-way homomorphic encryptions (i.e., each biometrics' feature vector) can determine matches or execute searches on encrypted data. Each biometrics' feature vector can then be stored and/or used in conjunction with respective classifications, for use in subsequent comparisons without fear of compromising the original biometric data. In various embodiments, the original biometric data is discarded responsive to generating the encrypted values. In another embodiment, the homomorphic encryption enables computations and comparisons on cypher text without decryption. This improves security over conventional approaches. Searching biometrics in the clear on any system, represents a significant security vulnerability. In various examples described herein, only the one-way encrypted biometric data is available on a given device. Various embodiments restrict execution to occur on encrypted biometrics for any matching or searching.

A computer-implemented method for determining entity characteristics. The method comprising: determining a numerical representation of each of a plurality of second entities, each numerical representation comprising a multi-dimensional vector characteristic of the respective second entity, wherein each second entity is associated with one or more first entities; reducing the dimensionality of the numerical representations to produce a plurality of reduced dimensionality numerical representations; and determining a specialisation of each of the plurality of first entities based on the respective reduced dimensionality numerical representations.

Methods and servers for storing data associated with users and digital items of a recommendation system having access to non-distributed and distributed storages. The server trains a model based for generating first user and item embeddings. The server stores (i) the first user embeddings in the non-distributed storage, and (ii) the first item embeddings in the distributed storage. The server re-trains the model for generating second user and item embeddings. The server stores (i) the second user embeddings in the non-distributed storage in addition to the first user embeddings, and (ii) second item embeddings in the distributed storage instead of the respective first item embeddings by replacing the respective first item embeddings. When the second item embeddings are stored on each node of the distributed storage, the server removes the first user embeddings associated with the first value from the non-distributed storage.

A vision-LiDAR fusion method and system based on deep canonical correlation analysis are provided. The method comprises: collecting RGB images and point cloud data of a road surface synchronously; extracting features of the RGB images to obtain RGB features; performing coordinate system conversion and rasterization on the point cloud data in turn, and then extracting features to obtain point cloud features; inputting point cloud features and RGB features into a pre-established and well-trained fusion model at the same time, to output feature-enhanced fused point cloud features, wherein the fusion model fuses RGB features to point cloud features by using correlation analysis and in combination with a deep neural network; and inputting the fused point cloud features into a pre-established object detection network to achieve object detection. A similarity calculation matrix is utilized to fuse two different modal features.

Systems and methods for determining a location based on image data are provided. A method can include receiving, by a computing system, a query image depicting a surrounding environment of a vehicle. The query image can be input into a machine-learned image embedding model and a machine-learned feature extraction model to obtain a query embedding and a query feature representation, respectively. The method can include identifying a subset of candidate embeddings that have embeddings similar to the query embedding. The method can include obtaining a respective feature representation for each image associated with the subset of candidate embeddings. The method can include determining a set of relative displacements between each image associated with the subset of candidate embeddings and the query image and determining a localized state of a vehicle based at least in part on the set of relative displacements.

A method for monitoring operation of a controller area network (CAN) comprising a plurality of nodes. The method comprises measuring a voltage associated with a CAN message transmitted on the network, determining a message signature in dependence on the measured voltage, and comparing the message signature with a node signature to determine the authenticity of the CAN message. One or more actions may be taken in dependence on the determined authenticity.

The present disclosure relates generally to machine learning. More particularly, the present disclosure relates to systems and methods that regularize neural networks by decorrelating neurons or other parameters of the neural networks during training of the neural networks promoting these parameter to innovate over one another.

Systems and methods for analyzing a customer deployment in a converged or hyper-converged infrastructure are disclosed. A machine learning model is trained based upon historical usage data of other customer deployments. A k-means clustering is performed to generate a prediction as to whether a deployment is configured for optimal failover. Recommendations to improve failover performance can also be generated.

Systems and methods are disclosed herein for automatically generating schemas for enterprise programs and initiatives and generating recommendations for adding parameters to those schemas. The parameters representing tasks, questions, people, milestones, and other suitable parameters to increase the likelihood that an enterprise program or an initiative (sometimes referred to as project) is successful.