Abstract: Methods, apparatus, programs, and storage media for enhancing speech signals based on a neural network are provided. The method includes receiving a speech signal and a corresponding reference speech signal. A first set of band features is extracted from the speech signal and a second set of band features is extracted from the reference speech signal, wherein each feature in the first set of band features corresponds to a frequency band of the speech signal and each feature in the second set of band features corresponds to a frequency band of the reference speech signal. A first mask is determined by inputting the third set of band features into a speech enhancement neural subnetwork. The low frequency mask part and the increased number of entries are combined into a second mask. A fourth set of band features is determined based on the second mask and the speech signal.

Abstract: Disclosed are a Mamba model-based lithium battery lifetime prediction method and system in the technical field of energy storage in renewable energy power systems, the method including: acquiring aging feature matrices of a plurality of lithium batteries and inputting the same into a Mamba model for training, extracting a hidden state matrix sequence of each of the lithium batteries based on the trained Mamba model, and obtaining a cycle count sequence corresponding to a target cycle life prediction point of each of the lithium batteries through norm calculation and variational mode decomposition; constructing a linear relationship model for lithium battery lifetime prediction based on cycle count sequences corresponding to target states of health of all of the lithium batteries and cycle count sequences corresponding to target cycle life prediction points of all of the lithium batteries.

Abstract: Methods and systems for training a model include training an encoder in an unsupervised fashion based on a backward latent flow between a reference frame and a driving frame taken from a same video. A diffusion model is trained that generates a video sequence responsive to an input image and a text condition, using the trained encoder to determine a latent flow sequence and occlusion map sequence of a labeled training video.

Abstract: The present disclosure relates to a visible light-catalyzed translucent concrete, and a preparation method and use thereof. The preparation method includes: extracting an iron oxide from a copper slag, mixing the iron oxide with TiO2 to obtain a photocatalyst, and then mixing the photocatalyst with an additive to obtain a photocatalytic slurry; preparing a concrete slurry using the copper slag after iron extraction as an aggregate; and pouring the photocatalytic slurry, the concrete slurry, and the photocatalytic slurry in sequence into a mold pre-laid with an optical fiber, to obtain the visible light-catalyzed translucent concrete. In the visible light-catalyzed translucent concrete, iron in the copper slag is used as a part of raw materials of the photocatalyst, and the copper slag after iron extraction is used as an aggregate to replace natural sand and gravel. This solves environmental pollutions caused by the copper slag and realizes resource utilization.

Abstract: Disclosed are a method, device, medium, and product for extracting feature maps from operational data of lithium batteries. The method includes: acquiring operational data of a lithium battery; selecting target operational data from the operational data to obtain a plurality of target data groups; sorting a plurality of operational data in each of the target data groups to obtain an initial matrix; performing mean-variance normalization processing on each column of the initial matrix to obtain a normalized matrix; calculating a covariance matrix of the normalized matrix; performing singular value decomposition on the covariance matrix to obtain an eigenvector matrix; calculating a dimensionality reduction matrix corresponding to each of the target data groups according to the eigenvector matrix; scaling the dimensionality reduction matrix to construct a three-dimensional feature map matrix; and converting the three-dimensional feature map matrix into an image to obtain a feature image of the lithium battery.

Abstract: The present disclosure discloses a parameter extraction method for an energy storage management system based on natural language interaction, and belongs to the technical field of energy storage for new energy power systems, comprising: dividing parameter types, collecting natural language interaction corpora, labeling parameters, constructing a word segmentation set, coding natural language interaction corpora, calculating prior probabilities and conditional probabilities, persisting the prior probabilities and the conditional probabilities, inputting new corpora, loading the prior probabilities and the conditional probabilities, calculating posterior probabilities according to a naive Bayes formula, and extracting parameters corresponding to a maximum value of the posterior probabilities. According to the present disclosure, parameters are extracted using a natural language interaction manner, which reduces a threshold of use of conventional energy storage management systems.

Abstract: Disclosed are an internal resistance consistency evaluation method and system for a dynamically reconfigurable battery module in the technical field of energy storage in power systems, where the method includes: calculating a current change rate by introducing a nominal capacity of battery cells; selecting a previous sampling time point of a sampling time point at which the current change rate is greater than or equal to a target current change rate as a target sampling time point; when the number of the target sampling time points is greater than a preset value, calculating an internal resistance of each of the battery cells at the target sampling time points according to the target sampling time points, and constructing an internal resistance matrix; utilizing the internal resistance matrix and applying an improved internal resistance consistency evaluation algorithm to calculate and obtain an internal resistance consistency result of the dynamically reconfigurable battery module.

Abstract: Disclosed herein is a system for implementing a management controller on a node, or network server, that is dedicated to monitoring the individual health of a plurality of accelerator modules configured on the node. Based on the monitored health, the management controller is configured to implement autonomous power cycle control of individual accelerator modules. The autonomous power cycle control is implemented without violating the requirements of standards established for accelerator modules (e.g., OPEN COMPUTE PROJECT requirements, PERIPHERAL COMPONENT INTERCONNECT EXPRESS (PCIe) interface requirements).

Abstract: Provided is an amorphous silicon thin-film transistor including an amorphous silicon semiconductor layer, a source electrode, and a drain electrode that are successively disposed on a base substrate. Ions doped by an ion implantation process are present in a region, proximal to the source electrode and the drain electrode, of the amorphous silicon semiconductor layer. A concentration of the ions in a surface region, proximal to the source electrode and the drain electrode, of the amorphous silicon semiconductor layer is greater than or equal to 5*10{circumflex over (?)}20 atoms/cc.

Abstract: An electrode balloon catheter and a high-voltage generation processing device are provided. The electrode balloon catheter includes a balloon, an inner catheter and a shock wave generation component. The balloon is disposed over the inner catheter and radially expands or collapses as a result of filling an inflation fluid therein or evacuating the inflation fluid therefrom. The shock wave generation component includes a flexible circuit layer and an electrode arrangement. The flexible circuit layer is disposed on the inner catheter, and the electrode arrangement is provided on the inner catheter to be located within the balloon. The electrode arrangement is connected to the flexible circuit layer and connected to a high-voltage generation processing device via the flexible circuit layer. The flexible circuit layer enables the electrode balloon catheter to have a reduced passage size closer to the size of a pre-dilation balloon.

Abstract: The present disclosure discloses a bus voltage modulation method for a series dynamic reconfiguration battery system, comprising: collecting a voltage value of each battery module, and arranging in a descending order according to magnitudes of the voltage values; selecting a battery module corresponding to the highest voltage value; determining whether the serial number i of the selected battery module is less than a total number N of series battery modules, if a determination result is YES, accumulating voltage values corresponding to the selected battery modules, and determining whether a voltage accumulation value of the battery modules is within a preset range; and repeating the above content until an output bus voltage meets requirements for the preset range. A dynamic reconfiguration battery system based on a software-defined technology of the present disclosure can reconstruct and output a desired bus voltage according to voltages of respective battery modules in real time.

Abstract: An embodiment for an improved method of automated version management for machine learning pipeline development is provided. The embodiment may compute a quality value of a target machine learning pipeline at a predetermined regular interval and automatically save updated versions of the target machine learning pipeline. The embodiment may extract a series of key features from detected versions of the target machine learning pipeline and cluster the detected versions of the target machine learning model pipeline by the extracted series of key features. The embodiment may identify a highest-quality version within each of a series of generated clusters. The embodiment may compute similarity scores for subsets of versions within each of the series of generated clusters. The embodiment may generate and output to a user, a visual representation of the series of generated clusters including version quality data and version similarity data.

Abstract: Methods and systems for training a model include performing spatial augmentation on an unlabeled input video to generate spatially augmented video. Temporal augmentation is performed on the input video to generate temporally augmented video. Predictions are generated, using a model that was pre-trained on a labeled dataset, for the unlabeled input video, the spatially augmented video, and the temporally augmented video. Parameters of the model are adapted using the predictions while enforcing temporal consistency, temporal consistency, and historical consistency. The model may be used for action recognition in a healthcare context, with recognition results being used for determining whether patients are performing a rehabilitation exercise correctly.

Abstract: An oral care device, an information display device thereof, and an information display method thereof are provided. The information display method is applied to the oral care device. The oral care device includes display areas respectively configured to display information of oral care positions. The information display method includes: determining a current oral care position of the oral care device and a current oral care duration corresponding to the current oral care position; when the current oral care duration is less than a predetermined duration corresponding to the current oral care position, controlling a display area corresponding to the current oral care position to display in a first manner; and when the current oral care duration is not less than the predetermined duration corresponding to the current oral care position, controlling the display area corresponding to the current oral care position to display in a second manner.

Abstract: The present disclose related to a system (1) and method for evaluating the performance of an audio processing scheme. The system (1) comprises an acoustic feature extractor (10A, 10B), configured to receive a plurality of segment pairs, each segment pair comprising a segment (101) and a processed segment (201). The acoustic feature extractor (10A, 10B) determines an acoustic feature associated with each segment and the system (1) further comprises an event detector (11), configured to receive the at least one acoustic feature of each segment (101, 201) and determine, for each segment pair and acoustic feature, if a difference between the acoustic feature of the segment and processed segment exceeds an event threshold. The system also comprises an event analyzer (12), configured to determine a performance metric based on each segment pair associated with a difference exceeding the event threshold.

Abstract: Described herein is a method of performing environment-aware processing of audio data for a mobile device. In particular, the method may comprise obtaining non-acoustic sensor information of the mobile device. The method may further comprise determining scene information indicative of an environment of the mobile device based on the non-acoustic sensor information. The method may yet further comprise performing audio processing of the audio data based on the determined scene information.

Abstract: The present application discloses a gold recovery system and recovery method thereof from a gold separation solution, where the recovery system includes an gold selective extraction system, a reduction system, a washing system, and an evaporation system; and the recovery method specifically includes the following steps: (1) primary extraction; (2) secondary extraction; (3) gold precipitation by reduction; (4) washing of gold powder; and (5) evaporation of waste solution. In the present application, by using chloride ions to recognize AuCl4?, the gold-rich solution is directly reduced to produce qualified gold powder product (99.995%); the gold-barren solution and the post-reduction solution are evaporated and concentrated, with the condensate reused to realize the recovery and regeneration of chloride ions, the recycling of the extractant and no generation of waste solution or waste gas in the production process.

Abstract: Disclosed is a battery aging assessment method based on multi-source and multi-scale high-dimensional state space modeling in the field of energy storage in renewable power systems. The method includes: acquiring a time series of each discharge process within a preset number of discharge cycles of a sample battery; determining a first state transition path and a second state transition path based on discharge parameters corresponding to the time series; establishing a benchmark working-state transition path; calculating multiple sample distances between the second state transition path and the benchmark working-state transition path; training a battery aging assessment model using the sample distances as input and corresponding target state-of-health values as output; calculating a target distance between a state transition path of a to-be-predicted target battery and the benchmark working-state transition path.

Abstract: Controlling application logging, including: identifying one or more relationships between a plurality of classes of an object oriented application; generating a tree structure comprised of a plurality of nodes and a plurality of branches, where each of the plurality of branches is comprised of at least one node from the plurality of nodes, based on the one or more identified relationships, where a first one or more nodes of the plurality of nodes correspond to one or more classes of the object-oriented application; assigning a weight to each of the plurality of branches, based at least in part on an identified relationship corresponding to each of the one or more nodes comprising each of the plurality of branches; and determining a logging level to each branch of the plurality of branches, based on the assigned weight for each branch of the plurality of branches.

Abstract: The present disclosure relates to a resource scheduling method and apparatus based on a distributed storage system, a medium, an electronic device, and a program product. The method includes: acquiring request retry failure information generated by a user side in a distributed storage cluster when a request retry fails, and storing the request retry failure information, wherein the request retry failure information is at least used to indicate a target storage resource on which the request retry fails; and in response to detecting that a target storage node in the distributed storage cluster has a heartbeat disconnection, scheduling, according to request retry failure information related to the target storage node, a target storage resource on which a request retry fails on the target storage node to another storage node in the distributed storage cluster that is in a normal operating state.