Abstract: Methods and apparatuses relating to processing neural networks are described. In one embodiment, an apparatus to process a neural network includes a plurality of fully connected layer chips coupled by an interconnect; a plurality of convolutional layer chips each coupled by an interconnect to a respective fully connected layer chip of the plurality of fully connected layer chips and each of the plurality of fully connected layer chips and the plurality of convolutional layer chips including an interconnect to couple each of a forward propagation compute intensive tile, a back propagation compute intensive tile, and a weight gradient compute intensive tile of a column of compute intensive tiles between a first memory intensive tile and a second memory intensive tile.

Abstract: The disclosed computer-implemented method for protecting a cloud computing device from malware may include (i) intercepting, at a computing device, a malicious attempt by the malware to (A) access sensitive information in an encrypted file stored on the computing device and (B) send the sensitive information to the cloud computing device and (ii) performing, responsive to the attempt to access the encrypted file, a security action. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for managing a need-to-know domain name system may include (i) intercepting, by an agent of the computing device, network traffic received on the computing device, (ii) generating, by the agent, a one-time password based on a unique identifier of the agent of the computing device, (iii) wrapping, by the agent, the network traffic with the one-time password, and (iv) pushing, by the agent, the wrapped network traffic to a cloud server using a local domain name system (DNS) of the agent of the computing device, wherein the local DNS comprises a private domain name unpublished in a global DNS. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method for protection of storage systems using decoy data may include identifying an original file comprising sensitive content to be protected against malicious access and protecting the sensitive content. Protecting the sensitive content may include (i) processing the original file to identify a structure of the original file and the sensitive content of the original file, (ii) generating a decoy file using the structure of the original file and using substitute content in a location corresponding to the sensitive content of the original file, and (iii) storing the decoy file with the original file. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: One embodiment of the present invention provides a system that facilitates selecting auction bids. During operation, the system receives a collection of bids, wherein each bid in the collection includes advertised performance criteria. For each bid in the collection, the system analyzes a past performance history for a bidder associated with the bid. Next, for each bid, the system determines an expected performance criteria based on the analyzed past performance history, and the advertised performance criteria. Finally, the system selects a winning bid from the collection of bids based on the expected performance criteria.

Abstract: One embodiment of the present invention provides a system that facilitates modeling satisfaction of a client based on an exponential-decay function. During operation, the system receives an identifier for a client. The system also receives a selection of components to use in modeling the satisfaction of the client. Note that a component is a performance metric, such as a goal, that has a quantifiable result. Next, for each selected component, the system determines a current satisfaction of the client by: determining an initial intensity for the satisfaction based on the component and an initial point in time; determining a decay constant for the client and the component; and using the decay constant in an exponential-decay function to decay the initial intensity to a current point in time, thereby obtaining the current satisfaction. Finally, the system combines the current satisfaction for each selected component to obtain an overall client satisfaction.

Abstract: A system that facilitates selecting an item from a cache based on a rank-order of the item picks a first item and a second item from the cache. Next, the system compares a rank-order of the first item with a rank-order of the second item. If the rank-order of the first item exceeds the rank-order of the second item, the system designates the first item as a winner of a first contest. However, if the rank-order of the first item is less than the rank-order of the second item, the system designates the second item as the winner of the first contest. Finally, the system selects the winner of the first contest as a selected item.

Abstract: One embodiment of the present invention provides a system that facilitates self-tuning traffic shaping without a central traffic manager. During operation, the system sends one or more requests from a client to a server over a specified period of time. Next, the system receives one or more responses to the requests from the server over the specified period of time. The system then determines a request-timeout percentage for the period of time. Finally, the system adjusts a rate at which requests are sent to the server based on the request-timeout percentage.

Abstract: A system and method for providing a filtering classloader within a Java-based or other application environment. In one embodiment, a filtering classloader is placed between two other classloaders, e.g. between the system classloader and the application classloader. The filtering classloader is provided with a set of patterns which throw a “class not found exception” if an application makes a request for that class or package. The exception leads the application classloader to believe (perhaps correctly) that there is no class with that name in the system classloader. The application classloader then loads the class itself. The class may or may not exist in the system classloader; the filtering classloader in essence overrides any retrieval from that classloader in favor of the application classloader, for certain specified patterns of request.

Abstract: An RFID edge server can associate with multiple RFID readers at a location, a service bus can receive RFID data from the RFID edge server and make the RFID data available to multiple services that consume the RFID data.

Abstract: One embodiment of the present invention provides a system for reducing reliance on a central data store while maintaining idempotency in a multi-client, multi-server environment. During operation, the system receives a request to perform a transaction from a client at a server, wherein the request includes a request identifier and an idempotence-risk flag. The system then determines if the idempotence-risk flag is set. If not, the system: processes the request at the server to produce a result, and stores the request in a request-cache local to the server. The system also stores the request in the central data store, and returns the result to the client.

Abstract: An RFID edge server using an application server allows for improvements in an RFID system.

Abstract: An RFID edge server can associate with multiple RFID readers at a location. The RFID edge server can include an application server using a Web Services Reliable Messaging to transfer RFID data.

Abstract: An RFID (Radio Frequency Identification) edge server associates with multiple RFID readers at a location. The RFID edge server includes an application server to run applications. The RFID edge server having TCP/IP socket connections with the multiple RFID readers. The RFID edge server having a fewer access threads for the RFID readers than there are socket connections.

Abstract: A system and method for using an automated process to identify bugs in software code. The system can be used to assist with the development of large and complex software products. File-level and/or line-level code coverage information is used to automatically trace-test failures to development changes within the product source code over a specified period of time. Information as to the health of a software product and the test criteria it passes or fails is compared at a first time, when the product may have satisfied all test criteria, with the health of the product at a second time when the same criteria may be failing. This information can then be used to narrow down and/or identify specific product failures to a particular change or set of changes in the software code, before any manual analysis need be done.

Abstract: An RFID edge server using an application server allows for improvements in an RFID system.

Abstract: A system and method for providing a filtering classloader within a Java-based or other application environment. In one embodiment, a filtering classloader is placed between two other classloaders, e.g. between the system classloader and the application classloader. The filtering classloader is provided with a set of patterns which throw a “class not found exception” if an application makes a request for that class or package. The exception leads the application classloader to believe (perhaps correctly) that there is no class with that name in the system classloader. The application classloader then loads the class itself. The class may or may not exist in the system classloader; the filtering classloader in essence overrides any retrieval from that classloader in favor of the application classloader, for certain specified patterns of request.

Abstract: An RFID edge server using an application server allows for improvements in an RFID system.

Abstract: Service oriented architecture can be used for an RFID system.

Abstract: Service oriented architecture can be used for an RFID system.