Abstract: The present application discloses a method and apparatus for obtaining a semantic label of a digital image. An implementation of the method includes: obtaining the digital image; looking up a semantic label model corresponding to the digital image, the semantic label model being used for representing correlation between digital images and semantic labels, and a semantic label being used for literally describing a digital image; and introducing the digital image into the semantic label model to obtain full-image recognition information and local recognition information corresponding to the digital image, and combining the full-image recognition information and the local recognition information to form a semantic label, the full-image recognition information being a summarized description of the digital image, and the local recognition information being a detailed description of the digital image.

Abstract: An embodiment for providing a secure virtual browsing environment includes creating a virtual browsing environment with a virtualized operating system sharing an operating system kernel of a supporting operating system and executing the browser application within the virtual browsing environment. Another embodiment includes receiving a website selection within a browser application, determining if the website selection corresponds to a secure bookmark, and creating a second virtual browsing environment and executing the browser application within the second virtual browsing environment to access the website selection when the website selection corresponds to a website specified as a secure bookmark.

Abstract: Described herein are systems and methods that address the task of learning novel visual concepts, and their interactions with other concepts, from a few images with sentence descriptions. Using linguistic context and visual features, embodiments are able to efficiently hypothesize the semantic meaning of new words and add them to model word dictionaries so that they can be used to describe images which contain these novel concepts. In the experiments, it was shown that the tested embodiments effectively learned novel visual concepts from a few examples without disturbing the previously learned concepts.

Abstract: A lens (100) is disclosed for a solid state lighting element (24). The lens comprises at least one light entry surface (110, 112) and a light exit surface (120) opposite the at least one light entry surface, the light exit surface comprising a regular pattern of microstructures (122) and a plurality of regular patterns of further microstructures (124), wherein each regular pattern of further microstructures is on a respective one of said microstructures. Such a lens (100) may achieve excellent colour mixing. A lighting device (10) including such a lens and a luminaire including such a lighting device (10) are also disclosed.

Abstract: Described herein are systems and methods that exploit hierarchical Recurrent Neural Networks (RNNs) to tackle the video captioning problem; that is, generating one or multiple sentences to describe a realistic video. Embodiments of the hierarchical framework comprise a sentence generator and a paragraph generator. In embodiments, the sentence generator produces one simple short sentence that describes a specific short video interval. In embodiments, it exploits both temporal- and spatial-attention mechanisms to selectively focus on visual elements during generation. In embodiments, the paragraph generator captures the inter-sentence dependency by taking as input the sentential embedding produced by the sentence generator, combining it with the paragraph history, and outputting the new initial state for the sentence generator.

Abstract: Described herein are systems and methods for generating and using attention-based deep learning architectures for visual question answering task (VQA) to automatically generate answers for image-related (still or video images) questions. To generate the correct answers, it is important for a model's attention to focus on the relevant regions of an image according to the question because different questions may ask about the attributes of different image regions. In embodiments, such question-guided attention is learned with a configurable convolutional neural network (ABC-CNN). Embodiments of the ABC-CNN models determine the attention maps by convolving image feature map with the configurable convolutional kernels determined by the questions semantics. In embodiments, the question-guided attention maps focus on the question-related regions and filters out noise in the unrelated regions.

Abstract: The present invention provides a tag processing method and device. In the embodiments of the present invention, through obtaining semantic characteristic data of a resource and then obtaining posterior probabilities of at least one tag sequence of the resource according to the semantic characteristic data of the resource, it is possible to select, based on the posterior probabilities, a tag sequence as a tag set for the resource, and thus realize the purpose of obtaining a plurality of tags for the resource.

Abstract: In one example a pulse measurement system, comprises at least one sensor positioned to collect pulse information at three separate locations along a body segment, a controller communicatively coupled to the at least one sensor to receive the pulse information from the at least one sensor, and a display coupled to the controller to present at least one characteristic of the pulse information detected at the three separate locations. Other examples may be described.

Abstract: A leaning grid device, also referred to as a fence assembly, for a miter saw includes at least one upper leaning grid which is a movable leaning grid having an upper support surface for a work piece and at least one lower leaning grid having a lower support surface for the work piece. The lower leaning grid is also a movable leaning grid and may move between two support positions. The lower support surface and the upper support surface are in a same plane when the lower leaning grid is located at the support positions and the lower support surface and the upper support surface are in different planes when the lower leaning grid is located at a position that is between the support positions.

Abstract: Presented herein are embodiments of a multimodal Recurrent Neural Network (m-RNN) model for generating novel image captions. In embodiments, it directly models the probability distribution of generating a word given a previous word or words and an image, and image captions are generated according to this distribution. In embodiments, the model comprises two sub-networks: a deep recurrent neural network for sentences and a deep convolutional network for images. In embodiments, these two sub-networks interact with each other in a multimodal layer to form the whole m-RNN model. The effectiveness of an embodiment of model was validated on four benchmark datasets, and it outperformed the state-of-the-art methods. In embodiments, the m-RNN model may also be applied to retrieval tasks for retrieving images or captions.

Abstract: Processor(s) for detecting malicious software. A hardware virtual machine monitor (HVMM) operates under a host OS. Container(s) initialized with network application template(s) operate under a guest OS VM. A detection module operates under the guest OS VM includes a trigger detection module, a logging module and a container command module. The trigger detection module monitors activity on container(s) for a trigger event. The logging module writes activity report(s) in response to trigger event(s). The container command module issues command(s) in response to trigger event(s). The command(s) include a container start, stop and revert commands. A virtual machine control console operates under the host OS and starts/stops the HVMM. A container control module operates under the guest OSVM and controls container(s) in response to the command(s). The server communication module sends activity report(s) to a central collection network appliance that maintains a repository of activities for infected devices.

Abstract: A lift device and a lift system are provided, which comprises: a carrying mechanism including a guiding hole; a lift pin passing through inside the guiding hole; a carrying base, which is connected to the lift pin, wherein when the carrying base is moved upward, the lift pin is moved upward and inside the guiding hole; and a magnetic control device, which is used to send a magnetic control signal, so as to move the lift pin downward. By the lift pin being moved downward and magnetically sucked, a bad process problem and a break risk can be avoided.

Abstract: A method including: receiving, at a first cell, radio resource information from a second cell, said first and second cells in at least one network controlled by two or more operators; receiving operator identity information and cell-type information at said first cell from said second cell; appending said operator identity information and said cell-type information to said radio resource information; and allocating radio resources in dependence on said information.

Abstract: Disclosed is a network frequency spectrum sharing method, a user terminal of the present network sending a network signal strength measurement report of the different network to a home base station; an operation management node of the present network, based on the network signal strength measurement report calculating a network overlapping level indicator of the different network, and obtaining network overlapping topology-based information of the different network; and the two networks exchanging the network overlapping topology-based information, and calculating a policy for allocating shared frequency spectrum resources therebetween. The present invention can enable different networks to fairly and reasonably share frequency spectrum resources within a specified range, thus increasing a frequency spectrum utilization rate.

Abstract: Disclosed is a wireless charging device, including: an oscillating module, a rectifier voltage reduction module, a power management module, and a Bluetooth communications module, wherein the Bluetooth communications module communicates with a charging socket to acquire a charging communications protocol, the oscillating module receives a signal from the charging socket and oscillates to generate AC, the rectifier voltage reduction module rectifies and reduces the voltage of the AC and then outputs a DC reference voltage, and the power management module reduces the reference voltage and then outputs a charging voltage to charge a battery.

Abstract: A method comprises determining in a base station of a first cell of a first network that there is an inactive second cell of a second, different network which is a candidate cell for a user equipment and causing an activation message to be sent to cause the activation of said inactive second cell.

Abstract: A lift device and a lift system are provided, which comprises: a carrying mechanism including a guiding hole; a lift pin passing through inside the guiding hole; a carrying base, which is connected to the lift pin, wherein when the carrying base is moved upward, the lift pin is moved upward and inside the guiding hole; and a magnetic control device, which is used to send a magnetic control signal, so as to move the lift pin downward. By the lift pin being moved downward and magnetically sucked, a bad process problem and a break risk can be avoided.

Abstract: An embodiment for providing a secure virtual browsing environment includes creating a virtual browsing environment with a virtualized operating system sharing an operating system kernel of a supporting operating system and executing the browser application within the virtual browsing environment. Another embodiment includes receiving a website selection within a browser application, determining if the website selection corresponds to a secure bookmark, and creating a second virtual browsing environment and executing the browser application within the second virtual browsing environment to access the website selection when the website selection corresponds to a website specified as a secure bookmark.

Abstract: A spectrum allocation method for a cellular network is presented. The method includes: abstracting a cellular network into an interference pattern using a graph theory-based spectrum distribution model; a network node establishing a stakeholder list to represent whether a spectrum can be shared with other network nodes; by obtaining stakeholder lists from other network nodes, calculating a fairness coefficient and determining a lower bound value for the number of spectrum resources which are used; meanwhile, each network node sending out a value related to the fairness coefficient, thereby enables calculate an upper bound value for the number of resources which can be obtained according to the obtained information, to effectively compete for free resources.

Abstract: Processor(s) for detecting malicious software. A hardware virtual machine monitor (HVMM) operates under a host OS. Container(s) initialized with network application template(s) operate under a guest OS VM. A detection module operates under the guest OS VM includes a trigger detection module, a logging module and a container command module. The trigger detection module monitors activity on container(s) for a trigger event. The logging module writes activity report(s) in response to trigger event(s). The container command module issues command(s) in response to trigger event(s). The command(s) include a container start, stop and revert commands. A virtual machine control console operates under the host OS and starts/stops the HVMM. A container control module operates under the guest OSVM and controls container(s) in response to the command(s). The server communication module sends activity report(s) to a central collection network appliance that maintains a repository of activities for infected devices.