Abstract: Embodiments of the present disclosure include systems and methods for compiling tensor operators for neural network models based on tensor tile configurations. A tensor expression for a neural network model is received. A tensor tile configuration for the tensor expression is determined based on specifications associated with a set of hardware devices. Based on the tensor tile configuration, a set of code for implementing the tensor expression for the neural network model is generated. The set of code is for execution on the set of hardware.

Abstract: Embodiments of the present disclosure include systems and methods for providing sparsity for neural network models based on sparsity attributes. A first neural network model definition is received. The first neural network model definition specifies a neural network model comprising a set of tensors and a set of sparsity attribute values for elements of a tensor in the set of tensors. The set of sparsity attribute values for the tensor are propagated to elements of a subset of the set of tensors to form a second neural network model definition. The neural network model is generated based on the second neural network model definition.

Abstract: According to implementations of the subject matter described herein, a solution is provided for pirated copy tracing based on a third party. In the solution, a report on a pirated copy of a digital content is received from a third party, wherein the report comprises first secret information for characterizing a first identification, time information and tracing information of the pirated copy. Subsequently, a request for verifying the report is received to determine whether the report is valid. When the report is determined as valid, a licensee associated with the report is marked as a first status to indicate that the pirated copy might be leaked by the licensee. Therefore, the pirated copy may be effectively traced based on third parties. The tracing information in the report can be hidden, and other third parties can therefore be prevented from using the tracing information for duplicate reports.

Abstract: According to implementations of the subject matter, a solution of dynamic management of computing resource is provided. In the solution, a first request for using a target number of computing resource in a set of computing resources is received, wherein at least one free computing resource of the set of computing resources is organized into at least one free resource group. When it is determined that a free matching resource group is absent from the first resource group and a free redundant resource group is present in at least one free resource group, the target number of computing resources are allocated for the first request by splitting the free redundant resource group, wherein the number of resources in the free redundant resource group is greater than the target number. Therefore, the dynamic allocation of computing resources is enabled.

Abstract: A computing device requests access to an application object from a remote storage system in order to locally execute application functionality without hosting application resources. An accessed object is associated with an intent in the storage system and locked. Locking an object in combination with an intent prevents computing devices that are not performing the intent from accessing the object. An intent defines one or more operations to be performed with the requested object, which are serialized as intent steps and stored in the storage system. Upon executing an intent step, the computing device stores a log entry at the storage system signifying the step's completion. A locked object remains locked until the log entries indicate every intent step as complete. Different computing devices can unlock a locked object by executing any incomplete steps of an intent associated with the locked object.

Abstract: Systems and methods described herein relate to secure, efficient, confidential, and/or outsourced blockchain networks, which can enable a group of mutually distrusting participants to securely share state and then agree on a linear history of operations on that shared state.

Abstract: Devices and methods for creating a totally ordered ledger of transaction may include assigning a ledger to the device, wherein the ledger includes transactions associated with a verifiable timestamp. The devices and methods may include providing a copy of the ledger to plurality of other devices in communication with the device. The devices and methods may include receiving copies of a plurality of other ledgers with other transactions associated with verifiable timestamps. The devices and methods may include generating an ordered ledger with an ordered list of transactions by performing a total order process that uses the verifiable timestamps of the transactions from the ledger and the verifiable timestamps of the other transactions from the copies of the plurality of other ledgers. The devices and methods may include executing the ordered list of transactions from the ordered ledger.

Abstract: According to implementations of the subject matter described herein, a solution for security management of a dataset is proposed. In this solution, a dataset comprising at least one record is obtained, a record of the at least one record at least comprising: a keyword for identifying the record; and a value corresponding to the keyword. Subsequently, a keyword index is created in a trusted execution environment on the basis of respective keywords of the at least one record. Here the keyword index describes a set of keywords of the at least one record. By means of the solution, the keyword index may be created for records in the dataset in the trusted execution environment, and based on the keyword index, the dataset may be managed in a more secure and reliable way so as to detect a possible anomaly in the dataset.

Abstract: A computing device requests access to an application object from a remote storage system in order to locally execute application functionality without hosting application resources. An accessed object is associated with an intent in the storage system and locked. Locking an object in combination with an intent prevents computing devices that are not performing the intent from accessing the object. An intent defines one or more operations to be performed with the requested object, which are serialized as intent steps and stored in the storage system. Upon executing an intent step, the computing device stores a log entry at the storage system signifying the step's completion. A locked object remains locked until the log entries indicate every intent step as complete. Different computing devices can unlock a locked object by executing any incomplete steps of an intent associated with the locked object.

Abstract: A computing device requests access to an application object from a remote storage system in order to locally execute application functionality without hosting application resources. An accessed object is associated with an intent in the storage system and locked. Locking an object in combination with an intent prevents computing devices that are not performing the intent from accessing the object. An intent defines one or more operations to be performed with the requested object, which are serialized as intent steps and stored in the storage system. Upon executing an intent step, the computing device stores a log entry at the storage system signifying the step's completion. A locked object remains locked until the log entries indicate every intent step as complete. Different computing devices can unlock a locked object by executing any incomplete steps of an intent associated with the locked object.

Abstract: This description relates to secure, efficient, confidential, and/or outsourced blockchain networks, which can enable a group of mutually distrusting participants to securely share state and then agree on a linear history of operations on that shared state.

Abstract: A computing device requests access to an application object from a remote storage system in order to locally execute application functionality without hosting application resources. An accessed object is associated with an intent in the storage system and locked. Locking an object in combination with an intent prevents computing devices that are not performing the intent from accessing the object. An intent defines one or more operations to be performed with the requested object, which are serialized as intent steps and stored in the storage system. Upon executing an intent step, the computing device stores a log entry at the storage system signifying the step's completion. A locked object remains locked until the log entries indicate every intent step as complete. Different computing devices can unlock a locked object by executing any incomplete steps of an intent associated with the locked object.

Abstract: A multi-party security protocol that incorporates biometric-based authentication and withstands attacks against any single party (e.g., mobile phone, cloud, or the user). The protocol involves the function split between mobile and cloud and the mechanisms to chain-hold the secrets. A key generation mechanisms binds secrets to a specific device or URL (uniform resource locator) by adding salt to a master credential. An inline CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) handling mechanism uses the same sensor modality as the authentication process, which not only improves the usability, but also facilitates the authentication process. This architecture further enhances existing overall system security (e.g., handling untrusted or compromised cloud service, phone being lost, impersonation, etc.) and also improves the usability by automatically handling the CAPTCHA.

Abstract: A machine-implemented method includes automatically determining that a host device is restarting from a disruptive stoppage of operations and that in-process write transactions by the host device to respective pages of non-volatile storage may have been interrupted. The method includes, in response to the determination, automatically scanning the non-volatile storage for all metadata-containing storage pages with respective identifications S(i) and having corresponding metadata relating each respective storage page S(i) to a corresponding data page P(j) and a corresponding version number V(k). The method includes automatically identifying scanned storage pages S(i) that have for their corresponding data page P(j) a most recent version number HV(k) and, in some cases, a secondmost recent version number.

Abstract: A cyclic commit protocol is used to store relationships between transactions and is used by the technology to determine whether a transaction is committed or not. The protocol allows creation of a cycle of transactions which can be used to recover the state of a storage device after a host failure by identifying the last committed version of intention records as committed or uncommitted based on the data stored in the physical pages.

Abstract: A continuous stream data (e.g., messages, tweets) is received by ingest nodes of a platform. The ingest nodes may analyze the data to create a transaction of graph updates, assign a sequence number to the transaction, and distribute the graph updates with the sequence number to graph nodes of the platform. The graph nodes may store graph updates from ingest nodes, and then the ingest nodes may report graph update progresses in a progress table. A snapshot may be taken based on the progress table, and then graph-mining computation may be implemented. Tolerate failures and decaying may be supported and incremental expansion may be allowed to cope with increasing update rates and computation needs.

Abstract: A cyclic commit protocol is used to store relationships between transactions and is used by the technology to determine whether a transaction is committed or not. The protocol allows creation of a cycle of transactions which can be used to recover the state of a storage device after a host failure by identifying the last committed version of intention records as committed or uncommitted based on the data stored in the physical pages.

Abstract: Data-parallel computation programs may be improved by, for example, determining the functional properties user defined functions (UDFs), eliminating unnecessary data-shuffling stages, and/or changing data-partition properties to cause desired data properties to appear after one or more user defined functions are applied.

Abstract: A continuous stream data (e.g., messages, tweets) is received by ingest nodes of a platform. The ingest nodes may analyze the data to create a transaction of graph updates, assign a sequence number to the transaction, and distribute the graph updates with the sequence number to graph nodes of the platform. The graph nodes may store graph updates from ingest nodes, and then the ingest nodes may report graph update progresses in a progress table. A snapshot may be taken based on the progress table, and then graph-mining computation may be implemented. Tolerate failures and decaying may be supported and incremental expansion may be allowed to cope with increasing update rates and computation needs.

Abstract: A multi-party security protocol that incorporates biometric-based authentication and withstands attacks against any single party (e.g., mobile phone, cloud, or the user). The protocol involves the function split between mobile and cloud and the mechanisms to chain-hold the secrets. A key generation mechanisms binds secrets to a specific device or URL (uniform resource locator) by adding salt to a master credential. An inline CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) handling mechanism uses the same sensor modality as the authentication process, which not only improves the usability, but also facilitates the authentication process. This architecture further enhances existing overall system security (e.g., handling untrusted or compromised cloud service, phone being lost, impersonation, etc.) and also improves the usability by automatically handling the CAPTCHA.