Abstract: An aeronautical aluminum alloy minimum-quantity-lubrication milling machining device includes a machine tool worktable and spindle connected with a machine tool power system. The spindle is connected with a tool holder that is fixed with a cutting tool. The machine tool worktable is provided with a machine tool fixture, the tool holder is connected with a minimum-quantity-lubrication mechanism, the machine tool fixture includes a fixture body that is fixedly provided with a limit block for contact with two adjacent side surfaces of a workpiece, the fixture body is provided with a plurality of clamping elements capable of pressing the workpiece against an upper surface of the fixture body, and a top of the clamping element is provided with a detection member for detecting a relative position between the clamping element and the spindle. The device can avoid interference and contact between a nozzle and the clamping element.

Abstract: An AI-assisted automatic labeling system and a method thereof are disclosed. The method comprises steps: selecting images from microscopic images as candidate images, using a pre-labeling module to automatically label cells in the candidate images, and dividing the labeled images into training data and verification data; using a training module and the training data to train a basic model; using a verification module to verify and modify the basic model, wherein the verification module respectively verifies at least one cell area and at least one background area of the verification data to converge the basic model and form an automatic labeling model; using the automatic labeling model to automatically label cells in redundant images of the microscopic images. The basic model trained by the present invention can use few labeled images to perform regressive training and verification and then automatically labels the redundant images accurately and efficiently.

Abstract: A semiconductor device according to the present disclosure includes a gate extension structure, a first source/drain feature and a second source/drain feature, a vertical stack of channel members extending between the first source/drain feature and the second source/drain feature along a direction, and a gate structure wrapping around each of the vertical stack of channel members. The gate extension structure is in direct contact with the first source/drain feature.

Abstract: Systems, methods, and computer program products for predicting user preference of items in an image process image data associated with a single image with a first branch of a neural network to produce an image embedding, the single image including a set of multiple items; process a user identifier of a user with a second branch of the neural network to produce a user embedding; concatenate the image embedding with the user embedding to produce a concatenated embedding; process the concatenated embedding with the neural network to produce a joint embedding; and generate a user preference score for the set of multiple items from the neural network based on the joint embedding, the user preference score including a prediction of whether the user prefers the set of multiple items.

Abstract: Provided is a method for implementing reinforcement learning by a neural network. The method may include performing, for each epoch of a first predetermined number of epochs, a second predetermined number of training iterations and a third predetermined number of testing iterations using a first neural network. The first neural network may include a first set of parameters, the training iterations may include a first set of hyperparameters, and the testing iterations may include a second set of hyperparameters. The testing iterations may be divided into segments, and each segment may include a fourth predetermined number of testing iterations. A first pattern may be determined based on at least one of the segments. At least one of the first set of hyperparameters or the second set of hyperparameters may be adjusted based on the pattern. A system and computer program product are also disclosed.

Abstract: A device includes a fin extending from a semiconductor substrate; a gate stack over the fin; a first spacer on a sidewall of the gate stack; a source/drain region in the fin adjacent the first spacer; an inter-layer dielectric layer (ILD) extending over the gate stack, the first spacer, and the source/drain region, the ILD having a first portion and a second portion, wherein the second portion of the ILD is closer to the gate stack than the first portion of the ILD; a contact plug extending through the ILD and contacting the source/drain region; a second spacer on a sidewall of the contact plug; and an air gap between the first spacer and the second spacer, wherein the first portion of the ILD extends across the air gap and physically contacts the second spacer, wherein the first portion of the ILD seals the air gap.

Abstract: A blackboard eraser cleaning device is provided and includes an outer shell (1). A cover (2) is arranged at the upper part of the outer shell (1), an inner shell (3) is arranged inside the outer shell (1). A clamping mechanism (19) is arranged at the upper end of the inner shell (3). A leaning station (8) is defined inside the clamping mechanism (19). A cylinder (9) is arranged around the clamping mechanism (19). A height adjustment mechanism (10) is arranged inside the outer shell (1). The height adjustment mechanism (10) includes a servo (101) arranged inside the outer shell (1) and located on a side of the inner shell (3). According to the present application, a variety of cleaning modes may be achieved to meet the actual usage of the blackboard eraser.

Abstract: A configuration for resource allocation for PDCCH with multi-relay based communication. The apparatus determines a transmission time of a signal comprising a sequence of a plurality of PDCCHs to a target node. At least one relay is between the base station and the target node. The apparatus transmits, to each of the at least one relay, a relay configuration comprising scheduling information for each of the at least one relay to relay the signal. The scheduling information is based on a determination of the transmission time of the signal to the target node.

Abstract: Active semiconductor devices in an integrated circuit are provided lateral electrical isolation by surrounding narrow deep trench isolation regions that are merged at shared portions of the narrow deep trench isolation regions. A wide deep trench isolation region laterally surrounds the merged narrow deep trench isolation regions.

Abstract: Embodiments of this specification disclose methods and systems for colocation of a plurality of business services. The method includes: for one of the business services, determining a target resource level corresponding to the business service; obtaining a target schedulable resource of each of one or more processing devices, where the target schedulable resource is a schedulable resource corresponding to the target resource level; and determining, based on whether the target schedulable resource satisfies a deployment condition corresponding to the business service, a method for deploying the business service on the one or more processing devices.

Abstract: A method of contrastive learning comprises: determining, based on a model construction criterion, a first encoder for a first modality and a second encoder for a second modality; constructing a first contrastive learning model, the first contrastive learning model comprising the first encoder and a third encoder for the second modality, and a model capacity of the third encoder being greater than a model capacity of the second encoder; performing pre-training of the first contrastive learning model based on a first training dataset for the first modality and the second modality; and providing the pre-trained first encoder in the pre-trained first contrastive learning model for a downstream task. Because only the model capacity of one encoder is increased in the pre-training stage, model performance may be improved without increasing model training overhead during downstream task fine-tuning and model running overhead during model application.

Abstract: The disclosure provides a method of active immunotherapy for a cancer patient, comprising administering vaccines against Globo series antigens (i.e., Globo H, SSEA-3 and SSEA-4). Specifically, the method comprises administering Globo H-CRM197 (OBI-833/821) in patients with cancer. The disclosure also provides a method of selecting a cancer patient who is suitable as treatment candidate for immunotherapy. Exemplary immune response can be characterized by reduction of the severity of disease, including but not limited to, prevention of disease, delay in onset of disease, decreased severity of symptoms, decreased morbidity and delayed mortality.

Abstract: Methods, apparatuses, a device, and a medium for training a contrastive learning model are provided. In a method, a plurality of sample sets for training the contrastive learning model are obtained, and the plurality of sample sets comprises a first sample set and a second sample set. A first target sample set is selected from the first sample set and the second sample set according to a predetermined rule. A first set of samples are determined based on the first target sample set according to a predefined batch size. The contrastive learning model is trained using the first set of samples. In this way, on the one hand, performance degradation of the contrastive learning model due to sample set bias may be avoided; on the other hand, a forgetting problem in the training process may be alleviated.

Abstract: A control method and a control system using the same as are provided. The method includes: extracting a corresponding component data from a component data stream based on an application protocol registered by each of a plurality of applications and encapsulating a corresponding application data packet for the application in response to receiving the component data stream in a current control period; transmitting the corresponding application data packet to each of the applications; obtaining control data reported by each of the applications according to data reporting timing corresponding to the application; and generating a component control instruction for controlling each of the components according to the control data reported by each of the applications in the current control period and the application protocol registered by the application. In this manner, the orderly control of a smart devices within a limited control period can be achieved.

Abstract: An array substrate includes a first pixel driving circuit, a second pixel driving circuit, and a third pixel driving circuit; and a first voltage supply line, a second voltage supply line, and a third voltage supply line configured to provide a constant voltage signal respectively to the first pixel driving circuit, the second pixel driving circuit, and the third pixel driving circuit. The first voltage signal line crosses over a second capacitor electrode in the first pixel driving circuit by a first overlapping area. The second voltage supply line crosses over a second capacitor electrode in the second pixel driving circuit by a second overlapping area. The third voltage supply line crosses over a second capacitor electrode in the third pixel driving circuit by a third overlapping area. The third overlapping area is greater than the first overlapping area, and is greater than the second overlapping area.

Abstract: A combined catalytic viscosity reducing system includes a catalyst slug, a heat generating system slug, a gas injection slug, and a water-soluble viscosity reducing system slug. The catalyst slug includes 10%-15% of azacarbene iron, 15%-30% of tert-butyl hydroperoxide, 2%-5% of phosphoric acid, 2%-5% of hydrogen donor, and 0.5%-1% of emulsifier agent, and the others are solvent. The heat generating system slug includes 10%-30% of NaNO2, 8%-25% of NH4Cl, and 3%-10% of acid initiator, and the others are water, totaling 100%. The water-soluble viscosity reducing system slug, according to mass percentage, includes 0.2%-0.5% of surfactant and 2%-10% of alkali, and the others are water. The combined catalytic viscosity reducing system can effectively reduce viscosity without injecting steam, and the viscosity reduction rate can reach 96.5%.

Abstract: Provided is a method for normalizing embeddings for cross-embedding alignment. The method may include applying mean centering to the at least one embedding set, applying spectral normalization to the at least one embedding set, and/or applying length normalization to the at least one embedding set. Spectral normalization may include decomposing the at least one embedding set, determining an average singular value of the at least one embedding set, determining a respective substitute singular value for each respective singular value of a diagonal matrix, and/or replacing the at least one embedding set with a product of the at least one embedding set, a right singular vector, and an inverse of the substitute diagonal matrix. The mean centering, spectral normalization, and/or length normalization may be iteratively repeated for a configurable number of iterations. A system and computer program product are also disclosed.

Abstract: Systems and methods are disclosed for accessing protected data. A computing device may have a secured stared storage accessible by two or more applications operating on the mobile device. The computing device may obtain a first token from an authorization service to verify user identity for a first application. The first token may be stored in the shared storage area, and be accessible to one or more applications sharing the storage space. In response to a user attempt to access a web service using a second application, the user identity may be verified using the first token. The authorization service may verify user credentials, and send a second token to the computing device. The second token may be a proxy ticket authorizing access and exchange of protected data between the second application and a web service. The second token may also be stored in the secure storage area.

Abstract: The present disclosure provides use of sweetpotato IbSAP15 gene in regulating leaf shape and flower shape of sweetpotato. Compared with the control, Ipomoea batatas cv. Xuzishu 8, IbSAP15-overexpressing lines have more deeply incised blades, dehiscent corollas, and higher ornamental value. In the present disclosure, overexpression of the sweetpotato IbSAP15 gene in sweetpotato can deepen leaf incisions, make trumpet-shaped corollas dehiscent, and change leaf shape and flower shape of sweetpotato. The present disclosure is suitable for developing sweetpotato germplasms with different leaf shape and flower shape and promotes ornamental flowering/foliage sweetpotato breeding.

Abstract: A display panel and a display device are provided. The display panel includes a display region and a non-display region. The non-display region includes a demultiplexing circuit, a fanout wire, and a clock signal line. An input terminal of the demultiplexing circuit is electrically connected to the fanout wire, an output terminal of the demultiplexing circuit is electrically connected to a data line in the display region, and a control terminal of the demultiplexing circuit is electrically connected to the clock signal line. The display panel further includes an isolation signal line including an isolation portion, and an orthographic projection of the isolation portion on a plane where a substrate is located is between an orthographic projection of the clock signal line on the plane where the substrate is located and an orthographic projection of the fanout wire on the plane where the substrate is located.