Abstract: Disclosed are an offset calibration method and circuit applied to a comparator array. The offset calibration circuit includes a global calibration voltage generation module, integrators, comparators, a global logic control circuit, and local logic control circuits; an entire comparator array shares the single global calibration voltage generation module, and each of the comparators only needs to introduce an integrator circuit, such that a large calibration range and a small calibration step are achieved, a long calibration cycle and a large circuit area are required, and the requirements for a low area are satisfied. The offset calibration method provided by the present disclosure is similar to a binary search algorithm, which is adopted to search for the offset voltage, the calibration step of 1 2 N ? Vref can be realized by only N cycles, and short calibration time is thus realized.

Abstract: Disclosed are an offset calibration method and circuit applied to a comparator array. The offset calibration circuit includes a global calibration voltage generation module, integrators, comparators, a global logic control circuit, and local logic control circuits; an entire comparator array shares the single global calibration voltage generation module, and each of the comparators only needs to introduce an integrator circuit, such that a large calibration range and a small calibration step are achieved, a long calibration cycle and a large circuit area are required, and the requirements for a low area are satisfied. The offset calibration method provided by the present disclosure is similar to a binary search algorithm, which is adopted to search for the offset voltage, the calibration step of 1 2 N ? Vref can be realized by only N cycles, and short calibration time is thus realized.

Abstract: A compressor and a moving scroll thereof includes a moving scroll wall integrally formed at a first side of a moving base plate, the moving scroll wall extending around an axis. A collection groove is disposed on an end face of the moving scroll wall, the end face being located at the end of the moving scroll wall that is opposite the moving base plate in the direction of the axis. A channel running in the direction of the axis is disposed between the collection groove and a second side, the channel and the collection groove being connected at a first port, and the channel running through the second side at a second port. The collection groove extends not less than 60 degrees and not more than 400 degrees around the axis from the first port to the collection end.

Abstract: A dynamic vortex disk for a compressor includes a dynamic substrate with a first side and a second side, and a dynamic vortex wall, which is integrally molded with the dynamic substrate, and extends around an axis of the dynamic substrate at the first side and away from the axis. The dynamic vortex wall further has an end surface opposite to the first side and away from the second side, and a groove deviating from a centerline of the dynamic vortex wall and formed on the end surface. The groove has a proximal end close to the axis of the dynamic substrate and an opposing distal end. An air supply channel is configured to extend along the axis through an interior of the dynamic vortex wall until it opens to the second side. The groove and the air supply channel are connected at the distal end.

Abstract: The disclosure relates to a method for regulating photocurrent of IGZO based on a two-dimensional black phosphorus material, and belongs to the field of semiconductor devices. The disclosure provides a method for preparing an IGZO-black phosphorus heterostructure by dry transfer technology, and by changing the contact mode between IGZO and black phosphorus, photocurrent response of the IGZO to different laser light wavelengths can be regulated. In the method, both a channel and electrodes of the IGZO are magnetron sputtered by means of masks, and the method has good repeatability and can realize preparation of large-area multi-devices. A black phosphorus sample is prepared by a mechanical exfoliation method, and has controllable thickness and size. A heterojunction is prepared by dry transfer technology, and the technology is easy to operate and highly controllable. The disclosure is beneficial to promote development of IGZO thin films in the micro-nano field and the semiconductor industry.

Abstract: A compressor and a moving scroll thereof includes a moving scroll wall integrally formed at a first side of a moving base plate, the moving scroll wall extending around an axis. A collection groove is disposed on an end face of the moving scroll wall, the end face being located at the end of the moving scroll wall that is opposite the moving base plate in the direction of the axis. A channel running in the direction of the axis is disposed between the collection groove and a second side, the channel and the collection groove being connected at a first port, and the channel running through the second side at a second port. The collection groove extends not less than 60 degrees and not more than 400 degrees around the axis from the first port to the collection end.

Abstract: An orbiting scroll plate includes a base including a first side and a second side, and a scroll wrap attached to the base at the first side and having a spiral cross-section. The scroll wrap extends from a first end to a second end, the first end is positioned close to the outer side of the base, and the second end is positioned close to the center of the base. A return channel connects to and is in communication with the first side and the second side, and a release channel connects to and is in communication with the first side and the second side and positioned closer to the outer side of the base than the return channel is. The scroll wrap and the base are integrally constructed. A scroll compressor containing the orbiting scroll plate is also described.

Abstract: The disclosure discloses a temperature compensation circuit and method for a neural network computing-in-memory array. Reference arrays sparsely inserted in the computing-in-memory array are adopted to provide a reference voltage for ADCs, so that an input voltage and a reference voltage of the ADCs have a same temperature coefficient. Finally, after analog-to-digital conversion by the ADC, the digital output of the ADC is not affected by the external temperature, thereby ensuring the operational precision of the neural network. According to the temperature compensation circuit of the disclosure, the reference arrays have the same structure as the computing-in-memory array. The insertion density of the reference arrays is related to the temperature field where the computing-in-memory arrays are located.

Abstract: The present invention provides a method for reconstructing a 3D object based on dynamic graph network, first, obtaining a plurality of feature vectors from 2D image I of an object; then, preparing input data: predefining an initial ellipsoid mesh, obtaining a feature input X by filling initial features and creating a relationship matrix A corresponding to the feature input X; then, inputting the feature input X and corresponding relationship matrix A to a dynamic graph network for integrating and deducing of each vertex's feature, thus new relationship matrix is obtained and used for the later graph convoluting, which improves the initial graph information and makes the initial graph information adapted to the mesh relation of the corresponding object, therefore the accuracy and the effect of 3D object reconstruction have been improved; last, regressing the position, thus the 3D structure of the object is deduced, and the 3D object reconstruction is completed.

Abstract: A method for machine reading comprehension includes: S1, obtaining a character-level indication vector of a question and a character-level indication vector of an article; S2, obtaining an encoded question vector and an encoded article vector; S3, obtaining an output P1 of a bidirectional attention model and an output P2 of a shared attention model; S4, obtaining an aggregated vector P3; S5, obtaining a text encoding vector P4; S6, obtaining global interaction information between words within the article; S7, obtaining a text vector P5 after using the self-attention model; S8, obtaining aggregated data P6 according to the text encoding vector P4 and the text vector P5; S9, obtaining a context vector of the article according to the aggregated data P6 and an unencoded article vector P; and S10, predicting an answer position according to the context vector of the article and the encoded question vector to complete the machine reading comprehension.

Abstract: A battery pack includes a box body including a mounting space. A cell array is disposed in the mounting space and includes multiple battery cores arranged in a first direction, and the first direction is parallel to a thickness direction of the battery cores. An adjustment shim set includes at least one adjustment shim and is disposed in the mounting space on a first side of the cell array in the first direction, and a first wedge-shaped member is disposed in the mounting space on a second side of the cell array in the first direction or a first side of the at least one adjustment shim away from the cell array, to press the cell array and the adjustment shim set. The number of adjustment shims of the adjustment shim set is configured to generate a pressure on the cell array in the mounting space to reach a pre-tightening force.

Abstract: The disclosure discloses a temperature compensation circuit and method for a neural network computing-in-memory array. Reference arrays sparsely inserted in the computing-in-memory array are adopted to provide a reference voltage for ADCs, so that an input voltage and a reference voltage of the ADCs have a same temperature coefficient. Finally, after analog-to-digital conversion by the ADC, the digital output of the ADC is not affected by the external temperature, thereby ensuring the operational precision of the neural network. According to the temperature compensation circuit of the disclosure, the reference arrays have the same structure as the computing-in-memory array. The insertion density of the reference arrays is related to the temperature field where the computing-in-memory arrays are located.

Abstract: The disclosure relates to a method for regulating photocurrent of IGZO based on a two-dimensional black phosphorus material, and belongs to the field of semiconductor devices. The disclosure provides a method for preparing an IGZO-black phosphorus heterostructure by dry transfer technology, and by changing the contact mode between IGZO and black phosphorus, photocurrent response of the IGZO to different laser light wavelengths can be regulated. In the method, both a channel and electrodes of the IGZO are magnetron sputtered by means of masks, and the method has good repeatability and can realize preparation of large-area multi-devices. A black phosphorus sample is prepared by a mechanical exfoliation method, and has controllable thickness and size. A heterojunction is prepared by dry transfer technology, and the technology is easy to operate and highly controllable. The disclosure is beneficial to promote development of IGZO thin films in the micro-nano field and the semiconductor industry.

Abstract: The present invention relates to microglial or myeloid expressed Alzheimer's disease associated (ME-AD) gene reporter constructs, cell lines and transgenic animals and their use for identifying agents that modulate the level or expression of ME-AD genes.

Abstract: Disclosed are a bailer-type long-shaft pump (100) and an associated pump station (200). The long-shaft pump comprises a bailer vehicle (110), a long shaft (120), a bearing (130), a transmission device (140) and an electric motor (150). One end of the long shaft (120) successively passes through the bearing (130) and the transmission device (140) so as to be in transmission connection with a drive shaft of the electric motor (150). The pump station (200) comprises a pump station foundation (210), several equipment rooms (231) transversely distributed at intervals and constructed on the pump station foundation (210), and two ends of the long shaft (120) are supported on side walls of two adjacent equipment rooms. The bailer-type long-shaft pump achieve large water flow and low lift, the pump station is space-saving, has shorter construction cycle, and reduces investment and operation costs.

Abstract: The present invention provides a method for reconstructing a 3D object based on dynamic graph network, first, obtaining a plurality of feature vectors from 2D image I of an object; then, preparing input data: predefining an initial ellipsoid mesh, obtaining a feature input X by filling initial features and creating a relationship matrix A corresponding to the feature input X; then, inputting the feature input X and corresponding relationship matrix A to a dynamic graph network for integrating and deducing of each vertex's feature, thus new relationship matrix is obtained and used for the later graph convoluting, which improves the initial graph information and makes the initial graph information adapted to the mesh relation of the corresponding object, therefore the accuracy and the effect of 3D object reconstruction have been improved; last, regressing the position, thus the 3D structure of the object is deduced, and the 3D object reconstruction is completed.

Abstract: An ESD protection device for bidirectional diode string triggering SCR structure belongs to the field of electro-static discharge of an integrated circuit. A deep N well is arranged on a P substrate, and a first P well, a first N well, a second P well and a second N well are successively arranged from left to right on a surface region of the deep N well. In a second N well region, a mask preparing plate is used to insert the P wells at intervals. The circumference of each P well is isolated by the N well. Each P well is respectively provided with a pair of P+ implantation region and N+ implantation region. The metal wire is connected with the implantation region, and a positive electrode and a negative electrode are led out from the metal wire for forward conduction and reverse conduction.

Abstract: A method for machine reading comprehension includes: S1, obtaining a character-level indication vector of a question and a character-level indication vector of an article; S2, obtaining an encoded question vector and an encoded article vector; S3, obtaining an output P1 of a bidirectional attention model and an output P2 of a shared attention model; S4, obtaining an aggregated vector P3; S5, obtaining a text encoding vector P4; S6, obtaining global interaction information between words within the article; S7, obtaining a text vector P5 after using the self-attention model; S8, obtaining aggregated data P6 according to the text encoding vector P4 and the text vector P5; S9, obtaining a context vector of the article according to the aggregated data P6 and an unencoded article vector P; and S10, predicting an answer position according to the context vector of the article and the encoded question vector to complete the machine reading comprehension.

Abstract: Disclosed are a bailer-type long-shaft pump and a pump station constructed using the long-shaft pump. The long-shaft pump comprises a bailer vehicle (110), a long shaft (120), a bearing (130), a transmission device (140) and an electric motor (150). The transmission device comprises a planetary gearbox (141), an orthogonal shaft gearbox (142) and a frequency converter (143); one end of the long shaft (120) successively passes through the bearing (130), the planetary gearbox (141), the orthogonal shaft gearbox (142) and the frequency converter (143), so as to be in transmission connection with a drive shaft of the electric motor (150); and the bailer vehicle (110) comprises a cylindrical base (112) and at least three blades (111) distributed on the surface of the cylindrical base (112) in an annular array radially extending from the center of the long shaft.