Abstract: Provided are systems for determining adversarial examples that include at least one processor to determine a first additional input from a plurality of additional inputs based on a proximity of the first additional input to an initial input, determine a second additional input from the plurality of additional inputs based on a proximity of the second additional input to the first additional input, generate a first vector embedding, a second vector embedding and a third vector embedding based on the second additional input, generate a first relational embedding, a second relational embedding, and a third relational embedding based on the third vector embedding and the first vector embedding, concatenate the first relational embedding, the second relational embedding, and the third relational embedding to provide a concatenated version, and determine whether the first input is an adversarial example based on the concatenated version. Methods and computer program products are also provided.

Abstract: The present application provides a flow battery testing device, including a housing and a testing body. The testing body includes a testing stack and a plurality of liquid path assemblies. The testing stack is provided with a plurality of liquid flow paths. An outer surface of the housing is provided with a bearing surface, and the testing stack is provided on the bearing surface. The plurality of liquid path assemblies are provided in the housing, and correspondingly cyclically communicate with the plurality of liquid flow paths one to one. By using a miniaturized flow battery testing device to replace an actual flow battery for relevant testing, it is possible to achieve test results equivalent to those of the actual flow battery, thus effectively reducing testing difficulty.

Abstract: An apparatus comprises: at least one optical source port providing an optical wave having a tunable spectral peak wavelength; one or more transmitting optical phased arrays (OPAs), each: coupled to the optical source port, wherein the OPAs form beams with a wavelength-controlled angular tuning range within a first plane and a phase-shift-controlled angular tuning range within a perpendicular plane; and receiving OPAs, each: coupled to a coherent receiver and configured to receive optical waves characterized by a wavelength-controlled angular tuning range within the first plane and a phase-shift-controlled angular tuning range within the perpendicular plane. At least two beams' wavelength-controlled angular tuning ranges are at least partially non-overlapping, and each of the receiving OPAs' wavelength-controlled angular tuning ranges at least one partially overlaps with at least one of the wavelength-controlled angular tuning ranges of the beams.

Abstract: This disclosure concerns a method of forming a covalent organic framework (COF) membrane, comprising forming a membrane substrate by impregnating a porous polymer with a pore-forming agent in order to form an impregnated polymer, at least partially carbonising the impregnated polymer at a temperature of about 150° C. to about 500° C. in order to form the membrane substrate, and interfacially polymerising amino monomers and acyl monomers on a surface of the membrane substrate in order to form the COF membrane. The membrane substrate is characterised by a crystallinity of about 10% to about 70% relative to the porous polymer. The disclosure also concerns the COF membrane thereof, and the use of the COF membrane in catalyst recovery.

Abstract: Techniques for suggesting media assets, the technique including: requesting a set of candidate media assets for a set of user media items based on a knowledge graph metadata network describing the set of user media items; receiving metadata for the set of candidate media assets; determining one or more sets of ranked media assets based on the received metadata; and outputting the determined one or more sets of ranked media assets.

Abstract: System, methods, and other embodiments described herein relate to implementing cooperative management strategies. In one embodiment, a method includes determining a severity score and a risk area. The method may then determine whether vehicles should belong to a first vehicle set, which may perform a cooperative vehicle response action, or a second vehicle set, in which individual response actions may be performed. The method may then generate a cooperative response action or an individual response action corresponding to the cooperative or individual vehicle sets, provided such vehicle sets are not empty.

Abstract: Systems, methods, and computer program products for determining long-range dependencies using a non-local graph neural network (GNN): receive a dataset comprising historical data; generate at least one layer of a graph neural network by generating graph convolutions to compute node embeddings for a plurality of nodes of the dataset, the graph convolutions generated by aggregating node data from a first node of the dataset and node data from at least one second node comprising a neighbor node of the first node; cluster the node embeddings to form a plurality of centroids; determine an attention operator for at least one node-centroid pairing, the at least one node-centroid pairing comprising the first node and a first centroid; and generate relational data corresponding to a relation between the first node and at least one third node comprising a non-neighbor node of the first node using the attention operator.

Abstract: A system and method is provided for implementing computer-controlled systems (e.g., vehicle safety and traffic management systems) based on compliance rates. A vehicle system and method are provided that estimate compliance rates in real-time, and have the capability to dynamically modify the computer-controlled instructions generated by a computer-controlled system (e.g., vehicle safety and traffic management systems) in a manner that improves compliance, and in-turn improves overall system performance. For example, a vehicle includes sensors receiving data from a plurality of communicative connected vehicles. The received data is associated with a real-time monitoring of traffic conditions for the vehicle. The vehicle also includes a controller device estimating a compliance rate in real-time based on the received data. The controller device of the vehicle also controls an automated operation of the vehicle using a modified computer-controlled instruction based on the estimated compliance rate.

Abstract: The invention belongs to the technical field of defense of new energy centralized control station networks, and particularly relates to a dynamic defense system and method of a new energy centralized control station network based on dynamic IP. By monitoring and statistics of the abnormal traffic, the present invention can determine the occurrence frequency and the occurrence interval of the abnormal traffic, further analyze and determine the defense state of a new energy centralized control station based on this, and execute the corresponding dynamic optimization solution according to different defense states, so that the IP address and the firewall of the new energy centralized control station can be dynamically adjusted to ensure the security performance of the new energy centralized control station, reduce the risk of external intrusion, effectively resist malicious network reconnaissance such as scanning attacks and ensure the stable operation of a new energy power generation system.

Abstract: Methods and systems are provided for synthesizing realistic time series data that may be used to better identify outliers within the synthesized realistic time series data. Noise can be introduced to a time domain representation of time series data and can introduce noise to a frequency domain representation of the time series data. Further, labeled anomalous points can be inserted into the time series data. The time series data may then be used for training a machine learning model to identify anomalies within new time series data.

Abstract: Described are a system, method, and computer program product for denoising sequential machine learning models. The method includes receiving data associated with a plurality of sequences and training a sequential machine learning model based on the data associated with the plurality of sequences to produce a trained sequential machine learning model. Training the sequential machine learning model includes denoising a plurality of sequential dependencies between items in the plurality of sequences using at least one trainable binary mask. The method also includes generating an output of the trained sequential machine learning model based on the denoised sequential dependencies. The method further includes generating a prediction of an item associated with a sequence of items based on the output of the trained sequential machine learning model.

Abstract: The present invention provides a fiber deployment method, storage medium, electronic device and system. The fiber deployment method includes the following steps: S10, traversing all single-mode fiber links and selecting a link for ultra-low loss fiber upgrade with the objective of minimizing the maximum number of frequency slots used in the whole network; and S20, when both an ultra-low loss fiber and a single-mode fiber satisfy the service demand, using preferentially spectrum resources in the single-mode fiber. The fiber deployment method of the present invention is simple and feasible. It allows a more efficient fiber upgrade strategy and more reasonable spectrum resource allocation and can make full use of the existing single-mode fibers in the elastic optical network, thereby allowing more efficient resource utilization.

Abstract: A developing cartridge includes: a cartridge containing a developer and including a first side wall and a second side wall positioned opposite to each other along a first direction; a power reception member configured to receive a driving power and positioned at the first side wall; a developing roller rotatable about a first axis extending in the first direction; and an activation member including a movable member and a detection protrusion positioned at the second side wall. The movable member rotates around a second axis in a second direction, the first direction and the second direction intersect with each other, and the detection protrusion moves as the movable member moves.

Abstract: Systems, methods, and computer program products that obtain a plurality of features associated with a plurality of samples and a plurality of labels for the plurality of samples; generate a plurality of first predictions for the plurality of samples with a first machine learning model; generate a plurality of second predictions for the plurality of samples with a second machine learning model; generate, based on the plurality of first predictions, the plurality of second predictions, the plurality of labels, and a plurality of groups of samples of the plurality of samples; determine, based on the plurality of groups of samples, a first success rate associated with the first machine learning model and a second success rate associated with the second machine learning model; and identify, based on the first success rate and the second success rate, a weak point in the machine learning first model or the second model.

Abstract: A switching board and a soldering method thereof are provided. The switching board includes a circuit board having a board body and a first metal ring, where the board body has a first surface, and the first metal ring is formed on the first surface and surrounds a periphery of the board body. The switching board further includes a first metal element disposed on the first surface and fastened to the first metal ring by soldering. The switching board has a high pressure resistance and a low leakage rate.

Abstract: A system and method that supports an integrated user interface experience with third party hosted applications selected and installed by a user within a user-installed native application of a user communication device.

Abstract: A method for dynamically assigning an inference request is disclosed. A method for dynamically assigning an inference request may include determining at least one model to process an inference request on a plurality of computing platforms, the plurality of computing platforms including at least one Central Processing Unit (CPU) and at least one Graphics Processing Unit (GPU), obtaining, with at least one processor, profile information of the at least one model, the profile information including measured characteristics of the at least one model, dynamically determining a selected computing platform from between the at least one CPU and the at least one GPU for responding to the inference request based on an optimized objective associated with a status of the computing platform and the profile information, and routing, with at least one processor, the inference request to the selected computing platform. A system and computer program product are also disclosed.

Abstract: A developing cartridge includes: a cartridge body for accommodating developers; a driving part at a first end in a first direction; a developing roller rotatable about a rotation axis extending along the first direction and positioned at a third end in a second direction; a storage medium for storing information; an electrical contact surface electrically connected to the storage medium and fixed relative to the cartridge body, and a lifting element configured to lift up the cartridge body in the third direction. The electrical contact surface is positioned at the first end in the first direction and is farther away from the developing roller relative to the driving part. The lifting element is farther away from the developing roller in the second direction away from the driving part.

Abstract: The present disclosure relates to a data processing method and apparatus based on a Merkel tree, the method optimizes the structure of the Merkel tree so that when generating a Merkel tree to be updated based on the original service data and inserting a new leaf node into the Merkel tree, no parent node is created for a leaf node without a sibling node and a non-leaf node without a sibling node; specially, when there is new service data, a corresponding new leaf node is created and inserted into the leaf node layer; if the new leaf node is located at an odd position in the leaf node layer, the non-leaf node layer does not need to be updated; if the new leaf node is located at an even position in the leaf node layer, a new non-leaf node is created and inserted into the non-leaf node layer.

Abstract: A network system for identifying a cable connection is provided. The network system includes a management server, a server device, and a storage device. The management server includes system-management software. The server device is connected to the management server. The server device includes a BMC configured to communicate with the system-management software of the management server. The storage device includes at least one cable port configured to receive a storage cable that connects the storage device to the server device. The cable port includes a non-volatile memory, an indicator light, and a I2C bus.