Abstract: Disclosed are a caffeic acid-based composite material and a preparation method thereof. The method includes: exposing a cyclodextrin metal-organic framework material prepared from ?-cyclodextrin to a solution of caffeic acid in a short-chain alcohol to obtain a mixed material; and incubating the mixed material, wherein during the incubating, the cyclodextrin metal-organic framework material is in dynamic contact with the solution of caffeic acid in a short-chain alcohol. The prepared composite material includes a cyclodextrin metal-organic framework material and caffeic acid loaded on the cyclodextrin metal-organic framework material, wherein the cyclodextrin metal-organic framework material is prepared from ?-cyclodextrin. The caffeic acid is loaded in an amount of 15-18% of a total mass of the caffeic acid-based composite material. The caffeic acid is located in a cavity of the cyclodextrin metal-organic framework material.

Abstract: A method may include forming a mask layer on top of a first dielectric layer formed on a first source/drain and a second source/drain, and creating an opening in the mask layer and the first dielectric layer that exposes portions of the first source/drain and the second source/drain. The method may include filling the opening with a metal layer that covers the exposed portions of the first source/drain and the second source/drain, and forming a gap in the metal layer to create a first metal contact and a second metal contact. The first metal contact may electrically couple to the first source/drain and the second metal contact may electrically couple to the second source/drain. The gap may separate the first metal contact from the second metal contact by less than nineteen nanometers.

Abstract: A method for compressing a neural network includes: obtaining a neural network including J operation layers; compressing a jth operation layer with Kj compression ratios to generate Kj operation branches; obtaining Kj weighting factors; replacing the jth operation layer with the Kj operation branches weighted by the Kj weighting factors to generate a replacement neural network; performing forward propagation to the replacement neural network, a weighted sum operation being performed on Kj operation results generated by the Kj operation branches with the Kj weighting factors and a result of the weighted sum operation being used as an output of the jth operation layer; performing backward propagation to the replacement neural network, updated values of the Kj weighting factors being calculated based on a model loss; and determining an operation branch corresponding to the maximum value of the updated values of the Kj weighting factors as a compressed jth operation layer.

Abstract: A method for compressing a neural network includes: obtaining a neural network including a plurality of parameters to be compressed; dividing the parameters into J blocks; compressing a jth block with Kj compression ratios to generate Kj operation branches; obtaining Kj weighting factors; replacing the jth block with the Kj operation branches weighted by the Kj weighting factors to generate a replacement neural network; performing forward propagation to the replacement neural network, a weighted sum operation being performed on Kj operation results generated by the Kj operation branches with the Kj weighting factors and a result of the operation being used as an output; performing backward propagation to the replacement neural network, updated values of the Kj weighting factors being calculated based on a model loss; and determining an operation branch corresponding to the maximum value of the updated values of the Kj weighting factors as a compressed jth block.

Abstract: The present application discloses a method and a device for pruning one or more convolutional layer in a neural network. The method includes: obtaining one target convolution layer from the one or more convolution layers in the neural network, the target convolution layer including C filters, each filter including K convolution kernels, and each convolution kernel including M rows and N columns of weight values, where C, K, M and N are positive integers greater than or equal to one; determining a number P of weight values to be pruned for each convolution kernel of the target convolution layer based on a number of weight values M×N in the convolution kernel and a target compression ratio, where P is a positive integer smaller than M×N; and setting P weight values with the smallest absolute values in each convolution kernel of the target convolution layer to zero to form a pruned convolution layer.

Abstract: A system and method for authenticating voice commands for a voice assistant includes an accelerometer device for recording vibrations corresponding to speech from a user. The recorded vibrations are compared to speech signals recorded by a microphone to determine if the speech signals originated from the user.

Abstract: A system and method for authenticating voice commands for a voice assistant includes an accelerometer device for recording vibrations corresponding to speech from a user. The recorded vibrations are compared to speech signals recorded by a microphone to determine if the speech signals originated from the user.

Abstract: Disclosed are ultrathin layers of group II-VI semiconductors, group II-VI semiconductor superlattice structures, photovoltaic devices incorporating the layers and superlattice structures and related methods. The superlattice structures comprise an ultrathin layer of a first group II-VI semiconductor alternating with an ultrathin layer of at least one additional semiconductor, e.g., a second group II-VI semiconductor, or a group IV semiconductor, or a group III-V semiconductor.

Abstract: The privacy threat of unregulated aggregation is addressed from a new perspective by monitoring, characterizing and reducing the underlying linkability across apps. This allows one to measure the potential threat of unregulated aggregation during runtime and promptly warn users of the associated risks. It was observed how real-world apps abuse OS-level information and IPCs to establish linkability. A practical countermeasure provides runtime monitoring and mediation of linkability across apps, introducing a new dimension to privacy protection on mobile device. An evaluation on real users has shown that proposed countermeasures are effective in reducing the linkability across apps and only incurs marginal overheads.

Abstract: A technique is presented for controlling behavior tracking by mobile apps executing on a mobile computing device. A set of profiles are managed by a privacy manager, where each profile represents a different isolated execution environment on the mobile computing device and defines a set of privacy rules governing an application while executing in an associated execution environment. Prior to launching the app, a user of the app is prompted by the privacy manager to select a profile for the app from the set of profiles. The privacy manager in turn applies the set of privacy rules associated with the selected profile to the app during execution of the app on the mobile computing device.

Abstract: Disclosed are ultrathin layers of group II-VI semiconductors, group II-VI semiconductor superlattice structures, photovoltaic devices incorporating the layers and superlattice structures and related methods. The superlattice structures comprise an ultrathin layer of a first group II-VI semiconductor alternating with an ultrathin layer of at least one additional semiconductor, e.g., a second group II-VI semiconductor, or a group IV semiconductor, or a group III-V semiconductor.

Abstract: The present disclosure disclosures a protecting system and method for privacy information in a mobile terminal. In the system, a crowdsourcing database stores security strategy information of applications. An interaction management module receives a recommendation request of security strategy sent by the mobile terminal. A recommendation algorithm module calculates according to the recommendation request of security strategy to obtain corresponding recommended security strategy. An interaction management module recommends the recommended security strategy to the privacy protecting module. A privacy protecting module notifies a user and assists the mobile terminal to set privacy protection strategy according to the recommended security strategy recommended by the interaction management module, and protects the privacy information in the mobile terminal according to the privacy protection strategy. The system can set privacy protection dynamically. In this way, user experience is improved.

Abstract: A technique is presented for controlling behavior tracking by mobile apps executing on a mobile computing device. A set of profiles are managed by a privacy manager, where each profile represents a different isolated execution environment on the mobile computing device and defines a set of privacy rules governing an application while executing in an associated execution environment. Prior to launching the app, a user of the app is prompted by the privacy manager to select a profile for the app from the set of profiles. The privacy manager in turn applies the set of privacy rules associated with the selected profile to the app during execution of the app on the mobile computing device.

Abstract: A data recovery method includes: by a first data node, obtaining a notification that a second data node fails; and storing specified data to a third data node, recording information of the specified data stored in the third data node in backup information stored in the first data node, and providing a metadata node and other data nodes storing the specified data with the information of the specified data stored in the third data node, where the specified data is data stored in the first and second data nodes. A data recovery method, two data nodes, and a distributed file system are also provided. In embodiments of the present invention, the data recovery is mainly performed among the data nodes, and the metadata node does not need to perform a lot of operations. Therefore, the load of the metadata node is reduced.