Abstract: Embodiments of the present disclosure provide a solution for data augmentation. A method includes: obtaining one or more candidate descriptions of an image with respect to a question associated with the image; determining a target description from the one or more candidate descriptions based on respective effectiveness metrics of the one or more candidate descriptions, an effectiveness metric of a candidate description indicating whether the candidate description is useful in answering the question; and constructing a training sample for a machine learning model, the training sample comprising the image, the question and the target description.

Abstract: Provided are a video processing method and apparatus, an electronic device, a storage medium and a program product. The method includes steps described below. A to-be-processed video is acquired, and a first face key point of a first target object in the to-be-processed video is recognized, where the first face key point corresponds to a mask key point in a three-dimensional human face mask effect; and the three-dimensional human face mask effect is added to video frames of the to-be-processed video to obtain a target video, where the mask key point in the three-dimensional human face mask effect moves with the corresponding first face key point.

Abstract: This application provides a fisheye image processing method, an electronic device, and a computer-readable storage medium. The method includes: acquiring a fisheye image; generating a grayscale image according to the fisheye image; generating a first gradient image in a horizontal direction and a second gradient image in a perpendicular direction; generating an azimuthal image according to the first gradient image and the second gradient image; and determining an effective imaging area according to the azimuthal image. The anti-distortion correction for the fisheye camera is performed in the process of shooting images according to the effective imaging area of the fisheye camera, and thus the imaging area of the fisheye camera is accurately calibrated when the images are shot by the fisheye camera.

Abstract: This application provides a fisheye image processing method, an electronic device, and a computer-readable storage medium. The fisheye method includes: extracting n sampling points in a fisheye image shot by a fisheye camera; determining n sampling points of a fisheye image; determining Cn3 sets of sampling points and corresponding circles; calculating errors of the sets of sampling points; calculating a first accumulated value of the errors and a second accumulated value of the errors of the sets of sampling points; constructing a nonlinear optimization equation; and determining an effective imaging area of the fisheye camera according to the nonlinear optimization equation. The anti-distortion correction for the fisheye camera is performed in the process of shooting images according to the effective imaging area of the fisheye camera.

Abstract: A retrieval-augmented neural transformer model with chunk cross-attention predicts a code review given a proposed source code change, represented as a code diff hunk, and a set of historical code review comments. The code diff hunk represents proposed edits to a source code snippet with its surrounding context that has not been changed. The historical code review comments are associated with code edits that are semantically similar to the proposed source code changes. The code diff hunk is partitioned into chunks which are used to find semantically similar historical code review comments. The set of historical code review comments is aggregated and used to guide the model in makings its predictions.

Abstract: This application discloses a calibration method for navigation of an UAV including a vector sensor. The calibration method includes: collecting, during a flight of the UAV, a current correction value and current data during a current measurement performed by the vector sensor; acquiring previous data during a previous measurement performed by the vector sensor; acquiring an adjustment quantity according to the current data and the previous data; acquiring a next correction value according to the current correction value and the adjustment quantity; and acquiring next original data during a next measurement performed by the vector sensor, acquiring next valid data according to the next original data and the next correction value, and controlling headings and postures of the UAV according to the next valid data. With the calibration method of this application, the next valid data Vk+1 more closely approximated to a true value can be obtained.

Abstract: A security protection method and device based on industrial Internet is provided. The disclosure relates to the technical field of network security and realize the isolation between the Internet and the industrial Internet platform intranet by deploying the exit firewall, and there is no restriction from the industrial Internet platform intranet to the Internet, and only necessary ports are opened from the Internet to the industrial Internet platform intranet. By deploying a regional firewall, isolation between the intranet core server and each of the secondary nodes is realized. The regional firewall is deployed on the wide area network router, and the second access policy of secondary nodes and intranet core servers is preset. After the second access policy is formulated, only the IP and service ports of specific hosts are opened, and all other accesses are prohibited.

Abstract: This application discloses a calibration method for navigation of an unmanned aerial vehicle (UAV), a non-transitory computer-readable storage medium and a UAV implementing the same. The calibration method includes: collecting, during a flight of the UAV, reference data during two measurements of a reference vector performed by a vector sensor; acquiring a zero-point offset M0 of the vector sensor according to the reference data; acquiring original data Rk of any vector measured by the vector sensor; acquiring valid data Vk according to the zero-point offset M0 and the original data Rk; and control headings and postures of the UAV according to the valid data Vk. With the calibration method in this application, the valid data Vk is defined as a vector data acquired after a zero-point error of the original data Rk is eliminated, which is more closely approximated to an actual value of a to-be-measured vector.

Abstract: Computing systems and technical methods that transform data structures and pierce opacity difficulties associated with complex machine learning algorithms are disclosed. Advances include a framework and techniques that include: i) global diagnostics; ii) locally interpretable models LIME-SUP-R and LIME-SUP-D; and iii) explainable neural networks.

Abstract: Provided are K-Ras mutant protein inhibitors of formula (I), a composition containing the same and the use thereof.

Abstract: A security protection method and device based on industrial Internet is provided. The disclosure relates to the technical field of network security and realize the isolation between the Internet and the industrial Internet platform intranet by deploying the exit firewall, and there is no restriction from the industrial Internet platform intranet to the Internet, and only necessary ports are opened from the Internet to the industrial Internet platform intranet. By deploying a regional firewall, isolation between the intranet core server and each of the secondary nodes is realized. The regional firewall is deployed on the wide area network router, and the second access policy of secondary nodes and intranet core servers is preset. After the second access policy is formulated, only the IP and service ports of specific hosts are opened, and all other accesses are prohibited.

Abstract: A Compound of formula (I) or (II), a stereoisomer thereof, an atropisomer, a pharmaceutically acceptable salt thereof, a pharmaceutically acceptable salt of the stereoisomer, a pharmaceutically acceptable salt of the atropisomer, a prodrug thereof, a deuterated molecule thereof or a conjugated form thereof is disclosed:

Abstract: Disclosed is an automatic crystal collection process, including: placing a crystal collection cylinder on a movable platform, detecting a length of a single crystal bar, and selecting a descending distance of the movable platform according to the length of the single crystal bar; raising a furnace sub-chamber to an upper limit, rotating the sub-chamber to a position above the crystal collection cylinder, and lowering the single crystal bar into the crystal collection cylinder; cutting off a seed crystal when the single crystal bar is lowered to a distance, and moving the single crystal bar with the descending of the movable platform until the single crystal bar is collected completely; raising the crystal collection cylinder along with the single crystal bar therein to a target position by the movable platform after the single crystal bar is collected completely, and finishing the crystal collection process.

Abstract: The present invention relates to N-macrocyclic amide compounds with structures shown in a general structural formula (I) and a general structural formula (II) or isomers, diastereomers, enantiomers, prodrugs and pharmaceutically acceptable salts of the N-macrocyclic amide compounds, wherein L, X, Y, Z, R1, A, B and C are respectively defined in the specification. The invention further relates to application of the N-macrocyclic amide compounds with the structure shown in the general structural formula (I) and the general structural formula (II) or isomers, diastereomers, enantiomers, prodrugs and pharmaceutically acceptable salts of the N-macrocyclic amide compounds as a medicament.

Abstract: A retrieval-augmented neural transformer model with chunk cross-attention predicts a code review given a proposed source code change, represented as a code diff hunk, and a set of historical code review comments. The code diff hunk represents proposed edits to a source code snippet with its surrounding context that has not been changed. The historical code review comments are associated with code edits that are semantically similar to the proposed source code changes. The code diff hunk is partitioned into chunks which are used to find semantically similar historical code review comments. The set of historical code review comments is aggregated and used to guide the model in makings its predictions.

Abstract: A constrained decoding technique incorporates token constraints into a beam search at each time step of a decoding process in order to generate viable candidate sequences that are syntactically and semantically correct. The token constraints identify source code tokens or sequences of tokens that should appear in a candidate sequence. The token constraints are generated from checking whether a token predicted at each decoding step is feasible for a partial solution based on the production rules of the grammar of the programming language, the syntactic correctness of a partial sequence, and/or static type correctness.

Abstract: A thermal management system includes a first heat exchange circuit, a second heat exchange circuit, a functional part capable of exchanging heat with the first heat exchange circuit, a heating module connecting with the second heat exchanging circuit and a valve connecting with the first and second heat exchanging circuits. The valve defines a first flow channel and a second flow channel prevented form communicating with each other inside the valve.

Abstract: The present application discloses a self-moving device, including a body, a chassis and a sensing unit. The body includes: a fluid storage apparatus used for storing a fluid; a cleaning liquid supply system used for supplying a cleaning liquid to a surface to be cleaned; and a dirty water recovery system used for recovering dirty water. The chassis is arranged at the lower part of the body for use in supporting the body. The sensing unit is arranged directly on the chassis.

Abstract: Provided herein is a compound of Formula (IB), a stereoisomer thereof, a pharmaceutically acceptable salt thereof, a pharmaceutically acceptable salt of the stereoisomer thereof, a prodrug thereof, a deuterated molecule thereof or a conjugated form thereof; composition containing the same and the use thereof.

Abstract: The embodiments provide a backplane and a power conversion method and apparatus. The power conversion apparatus includes a first power module, a second power module, and a controller. Each power module includes a signal processor and a corresponding power converter. The signal processor in the first power module sends a generated first carrier signal to the signal processor in the second power module. The signal processor in the second power module determines a second carrier signal by using a period of the first carrier signal, to drive the power converter in the first power module by using the first carrier signal, and drive the power converter in the second power module by using the second carrier signal. This improves operating efficiency of a device in a power conversion process and reduces a power loss.