Abstract: A magnetic random access memory (MRAM) structure is provided. The MRAM structure includes a first write electrode, a first magnetic tunnel junction (MTJ) stack, a voltage control electrode, a second MTJ stack, and a second write electrode. The first MTJ stack includes a first free layer disposed on the first write electrode, a first tunnel barrier layer disposed on the first free layer, and a first fixed layer disposed on the first tunnel barrier layer. The voltage control electrode is disposed on the first MTJ stack. The second MTJ stack includes a second fixed layer disposed on the voltage control electrode, a second tunnel barrier layer disposed on the second fixed layer, and a second free layer disposed on the second tunnel barrier layer. The second write electrode is disposed on the second MTJ stack.

Abstract: A method performed by at least one processor of an encoder comprises: performing, on an input 3D mesh, a symmetry detection process to estimate a partition plane. The method further comprises partitioning the input 3D mesh into a first side and a second side based on the partition plane, where the first side is opposite to the second side, and where a first vertex on the first side is symmetric to a second vertex on the second side. The method further comprises quantizing the first vertex and the second vertex in pairs to reduce a quantization error and symmetry prediction error associated with the first vertex and the second vertex.

Abstract: An image recognition method using reduced processes of calculation acquires training data and trains parameters of fully connected layers of a neural network model with the data. The neural network model is an image recognition model, the training of the parameters of fully connected searches for nodes of at least one connected layer and divides the nodes in a same layer of the into multiple groups according to a first preset rule. A feature node in each of the multiple groups is determined according to a second preset rule and feature node information in each group is used as parameter information to complete the training of the fully connected layer. An image is acquired and input in response to receiving an image recognition command, and an image recognition result is output. A computing device applying the method is also disclosed.

Abstract: Method, apparatus, and system for merging multiple attribute maps for mesh compression may be provided. The process may include obtaining multiple attribute maps associated with a mesh including one or more texture maps. The multiple attribute maps may be concatenated into a single concatenated map and concatenated UV coordinates for each of the multiple attribute maps may be generated based on re-computing original UV coordinates of each of the multiple attribute maps within the single concatenated map.

Abstract: An in-plane magnetized spin-orbit magnetic device is provided. The in-plane magnetized spin-orbit magnetic device includes a heavy metal layer, an upper electrode and a magnetic tunnel junction. The magnetic tunnel junction is disposed between the heavy metal layer and the upper electrode. The magnetic tunnel junction includes a free layer and a pinned layer. The free layer is disposed on the heavy metal layer, and the free layer has a first film plane area. The pinned layer is disposed on the free layer, and the pinned layer has a second film plane area. There is a preset angle between a long axis direction of a film plane shape of the free layer and a long axis direction of a film plane shape of the pinned layer, and the first film plane area is larger than the second film plane area.

Abstract: An in-plane magnetized spin-orbit magnetic device is provided. The in-plane magnetized spin-orbit magnetic device includes a heavy metal layer, an antiferromagnetic layer, and a magnetic tunnel junction. The antiferromagnetic layer is disposed on the heavy metal layer, and the magnetic tunnel junction is disposed on the antiferromagnetic layer. The magnetic tunnel junction includes a free layer, a barrier layer, and a pinned layer. The barrier layer is disposed on the free layer, and the pinned layer is disposed on the barrier layer. A film surface shape of the free layer is a rounded rectangle.

Abstract: A method for optimizing operation of convolution neural network outputs an input matrix of an input image. The method further slides on the input matrix according to preset convolution kernels to perform dot products to output first output matrixes on a convolution computing layer. Nonlinear mapping of the first output matrixes is performed according to a preset activation functions on an activation layer to output a second output matrixes. The method does not perform bias operations on the convolution computing layer. The results of the dot products performed on the convolution computing layer are output to the activation layer and bias operations are performed on the activation layer. The method reduces amount of calculation whilst not reducing accuracy of processing. An electronic device, a convolution neural network, and a non-transitory storage medium are also disclosed.

Abstract: An in-plane magnetized spin-orbit magnetic device is provided. The in-plane magnetized spin-orbit magnetic device includes a heavy metal layer, an upper electrode and a magnetic tunnel junction. The magnetic tunnel junction is disposed between the heavy metal layer and the upper electrode. The magnetic tunnel junction includes a free layer and a pinned layer. The free layer is disposed on the heavy metal layer, and the free layer has a first film plane area. The pinned layer is disposed on the free layer, and the pinned layer has a second film plane area. There is a preset angle between a long axis direction of a film plane shape of the free layer and a long axis direction of a film plane shape of the pinned layer, and the first film plane area is larger than the second film plane area.

Abstract: The disclosure provides a method for verifying training data, a training system, and a computer program produce. The method includes: receiving a labelled result with a plurality of bounding regions, wherein the labelled result corresponds to an image, the bounding regions are labelled by a plurality of annotators, the annotators comprises a first annotator and a second annotator, and the bounding region comprises a first bounding region labelled by the first annotator and a second bounding region labelled by the second annotator; and determining the first bounding region and the second bounding region respectively corresponds to different two target objects or corresponds to one target object according to a similarity between the first bounding region and the second bounding region.

Abstract: An image recognition method using reduced processes of calculation acquires training data and trains parameters of fully connected layers of a neural network model with the data. The neural network model is an image recognition model, the training of the parameters of fully connected searches for nodes of at least one connected layer and divides the nodes in a same layer of the into multiple groups according to a first preset rule. A feature node in each of the multiple groups is determined according to a second preset rule and feature node information in each group is used as parameter information to complete the training of the fully connected layer. An image is acquired and input in response to receiving an image recognition command, and an image recognition result is output. A computing device applying the method is also disclosed.

Abstract: A flexible photovoltaic module is provided. The flexible photovoltaic module includes a front plate, a solar cell and a back plate. The front plate, the solar cell and the back plate are arranged from top to bottom. The solar cell includes a polyethylene terephthalate (PET) based film arranged on a side of the solar cell close to the front plate. The PET based film includes a first adhesive layer arranged on a side of the PET based film close to the front plate. The back plate includes a second adhesive layer arranged on a side of the back plate close to the solar cell.

Abstract: The present disclosure discloses a reliability test device for flexible photovoltaic module, this reliability test device for flexible photovoltaic module comprises: an environment test box, a temperature acquisition means and a placing rack placed inside an environment test box; the placing rack comprises at least one set of oppositely provided vertical frames, and a plurality of carriers horizontally stacked between the vertical frames, the carriers and the vertical frames are fixedly connected; a first gap is provided between the carriers, and a plurality of gas holes are provided on the carriers; the flexible photovoltaic module is horizontally placed on the carrier, and the temperature acquisition means is fixedly provided on a surface of the flexible photovoltaic module.

Abstract: A perpendicularly magnetized spin-orbit magnetic device including a heavy metal layer, a magnetic tunnel junction, a first antiferromagnetic layer, a first block layer and a first stray field applying layer is provided. The magnetic tunnel junction is disposed on the heavy metal layer. The first block layer is disposed between the magnetic tunnel junction and the first antiferromagnetic layer. The first stray field applying layer is disposed between the first antiferromagnetic layer and the first block layer. The magnetic tunnel junction comprises a free layer, a tunneling barrier layer, and pinned layer. The tunneling barrier layer is disposed on the free layer. The pinned layer is disposed on the tunneling barrier layer. A film plane area of the free layer is greater than a film plane area of the tunneling barrier layer and a film plane area of the pinned layer.

Abstract: The disclosure provides a method for verifying training data, a training system, and a computer program produce. The method includes: receiving a labelled result with a plurality of bounding regions, wherein the labelled result corresponds to an image, the bounding regions are labelled by a plurality of annotators, the annotators comprises a first annotator and a second annotator, and the bounding region comprises a first bounding region labelled by the first annotator and a second bounding region labelled by the second annotator; and determining the first bounding region and the second bounding region respectively corresponds to different two target objects or corresponds to one target object according to a similarity between the first bounding region and the second bounding region.

Abstract: The disclosure provides a method for verifying training data, a training system, and a computer program produce. The method includes: providing a plurality of raw data to a plurality of annotators; retrieving a plurality of labelled results, wherein the labelled results includes a plurality of labelled data, and the labelled data are generated by the annotators via labelling the raw data; determining a plurality of consistencies by comparing the labelled results, and accordingly determining whether the labelled results are valid for training an artificial intelligence machine; in response to determining that the labelled results are valid, determining at least a specific part of the labelled results are valid for training the artificial intelligence machine.

Abstract: A perpendicularly magnetized spin-orbit magnetic device including a heavy metal layer, a magnetic tunnel junction, a first antiferromagnetic layer, a first block layer and a first stray field applying layer is provided. The magnetic tunnel junction is disposed on the heavy metal layer. The first block layer is disposed between the magnetic tunnel junction and the first antiferromagnetic layer. The first stray field applying layer is disposed between the first antiferromagnetic layer and the first block layer. The magnetic tunnel junction comprises a free layer, a tunneling barrier layer, and pinned layer. The tunneling barrier layer is disposed on the free layer. The pinned layer is disposed on the tunneling barrier layer. A film plane area of the free layer is greater than a film plane area of the tunneling barrier layer and a film plane area of the pinned layer.

Abstract: A perpendicularly magnetized spin-orbit magnetic device including a heavy metal layer, a magnetic tunnel junction, a first antiferromagnetic layer, a first block layer and a first stray field applying layer is provided. The magnetic tunnel junction is disposed on the heavy metal layer. The first block layer is disposed between the magnetic tunnel junction and the first antiferromagnetic layer. The first stray field applying layer is disposed between the first antiferromagnetic layer and the first block layer. The first stray field applying layer provides a stray magnetic field parallel to a film plane. The first antiferromagnetic layer contacts the first stray field applying layer to define the direction of the magnetic moment in the first stray field applying layer.

Abstract: The present disclosure discloses a method for preparing an insulating layer of a flexible photovoltaic module. The method includes: placing the flexible photovoltaic module on a spraying device, and using a shielding component to shield both side surfaces of the flexible photovoltaic module; performing a first spraying on the sidewall to be insulated of the flexible photovoltaic module to form an insulating bottom layer; performing a second spraying on the insulating bottom layer to form an insulating top layer such that the insulating bottom layer and the insulating top layer constitute an insulating layer; and drying the insulating layer.

Abstract: A perpendicularly magnetized spin-orbit magnetic device including a heavy metal layer, a magnetic tunnel junction, a first antiferromagnetic layer, a first block layer and a first stray field applying layer is provided. The magnetic tunnel junction is disposed on the heavy metal layer. The first block layer is disposed between the magnetic tunnel junction and the first antiferromagnetic layer. The first stray field applying layer is disposed between the first antiferromagnetic layer and the first block layer. The first stray field applying layer provides a stray magnetic field parallel to a film plane. The first antiferromagnetic layer contacts the first stray field applying layer to define the direction of the magnetic moment in the first stray field applying layer.

Abstract: The present disclosure discloses an outdoor test device for a variable-angle photovoltaic module including a base, a turntable and a storage table, wherein a caster assembly is fixedly arranged on the base, the turntable is rotatably connected with the base, a beam angle pole is fixedly arranged on the turntable, the beam angle pole is perpendicular to the turntable, and the storage table is fixedly connected with the base.