Abstract: Disclosed are an optimization method and apparatus for compiling computation graph. The optimization method includes the following steps: step S1: converting a computation graph into an intermediate representation; step S2: analyzing a dependency relationship; step S3: constructing a work stack; step S4: performing initialization to achieve a nonactivated state; step S5: popping out stack top node elements, and updating an input node set in a current round of iteration; step S6: adding the stack top node elements that depend on step S5 to a stack top position in sequence until the work stack is empty; step S7: implementing an intermediate representation in a fixed node state using a bit vector; and step S8: allocating registers for effective tensor variables contained in nodes of the intermediate representation in the fixed node state.

Abstract: The present disclosure provides an organic light-emitting diode display substrate. The organic light-emitting diode display substrate includes: a light-emitting layer, a light modulation layer, and a color conversion layer, in which the light-emitting layer is configured to emit first color light, the light modulation layer and the color conversion layer are arranged on different light-exiting paths of the light-emitting layer, the color conversion layer is configured to convert first color light into second color light and third color light, and the light modulation layer is configured to modulate an emergent direction of first color light.

Abstract: Disclosed are an instruction execution method and apparatus for graph computation. The method includes the following steps: S1: sending operators of each node in a computational graph used for neural network computation to an operator interpreter; S2: building, by the operator interpreter, instructions in operation; S3: defining an instruction dependency relationship; S4: building an instruction dependency relationship graph; S5: building a topological order of parallel instructions; S6: scheduling the parallel instructions to hardware resources; S7: building shortest schedules for the parallel instructions: the shortest time required to execute the parallel instructions under the condition of limited hardware resources; and S8: releasing the completed instructions.

Abstract: A memory optimization method includes: compiling a neural network into a computational graph for neural network computation on a computer; transforming the computational graph into a topological graph; constructing a life cycle relationship graph of tensor variables in the computational graph; and analyzing a life cycle relationship among tensor variables in a node of the computational graph; iteratively merging those tensor variables connected by lines of the second type and caching into a memory any tensor variable that goes beyond a number of idle registers and is not allocated to a register, until all tensor variables that go beyond the number of the idle registers and are not allocated to registers are cached into the memory; caching any node of the life cycle relationship graph with a degree smaller than a number of registers into a stack.

Abstract: The disclosure discloses an intermediate representation method for compiling computation graphs, including: step 1: compiling a neural network into a computation graph for neural network computation; step 2: constructing a node for each tensor variable in the computation graph; step 3: associating the node representing the tensor variable in the computation graph to a set of pointers to the tensor variable; step 4: analyzing constraint relationships between the tensor variables in the computation graph; step 5: iteratively constructing a topological graph of the intermediate representation based on the constraint relationships between the tensor variables in the computation graph; and step 6: analyzing the tensor variables with different aliases pointing to a same memory location based on the intermediate representation, and allocating a register for the tensor variables with different aliases.

Abstract: Disclosed are a memory optimization method and device oriented to neural network computing. The memory optimization method oriented to neural network computing includes the following steps: step S1: reconstructing a computation graph into a topological structure computation graph; step S2: constructing a life cycle interval about tensor variables; step S3: constructing a scanning line about the life cycle interval; step S4: allocating the tensor variables to idle registers; step S5: allocating to tensor variables exceeding the required number of registers; step S6: allocating registers allocated in the expired life cycle interval to tensor variables exceeding the required number of registers; and step S7: adding tensor variables transferred to a memory back to the life cycle interval in an activated state, and allocating idle registers for the tensor variables. According to the present disclosure, the memory of a data flow of a computation graph for neural network computing is optimized.

Abstract: Disclosed herein are an adhesion blocking element, a three-dimensional printing device and a three-dimensional printing method. The adhesion blocking element comprises: one light-transmittable main body comprising a first surface and a second surface which are disposed opposite to each other, and side faces connecting the first surface and the second surface; and a plurality of microstructures arranged on the main body, wherein each microstructure has one cavity formed in the main body and one first open face which is arranged on the first surface of the main body and communicated to the cavity. The present invention decreases the adhesion between the adhesion blocking element and the cured layer by improving the structure of the adhesion blocking element itself, and eliminates the negative pressure adsorption between the cured layer and the adhesion blocking element, so that it is easier to peel the adhesion blocking element off from the cured layer.

Abstract: Embodiments of the present disclosure provide a line laser module and an autonomous mobile device. The line laser module includes a fixed base, and a camera and a line laser emitter arranged on the fixed base. The line laser emitter is provided at one or more sides of the camera, and configured to emit a laser with a linear projection. The camera is configured to operate in conjunction with the line laser emitter, and to capture an environmental image. An infrared filter is arranged in front of the camera, and configured to allow only infrared light to enter the camera. The autonomous mobile device includes an infrared flashlight. The camera is configured to capture, at different time points, a first environmental image for distance measurement and a second environmental image for object identification.

Abstract: The present disclosure discloses a method for execution of a computational graph in a neural network model and an apparatus thereof, including: creating task execution bodies on a native machine according to a physical computational graph compiled and generated by a deep learning framework, and designing a solution for allocating a plurality of idle memory blocks to each task execution body, so that the entire computational graph participates in deep learning training tasks of different batches of data in a pipelining and parallelizing manner.

Abstract: Disclosed are a data flow method and apparatus for neural network computation. The data flow method for neural network computation includes initializing the lifecycle of a variable in a computational graph; and defining a propagation rule for a variable in use to flow through a node. A definition of the variable is produced at a precursor node of the node, such that an input set of valid variables flowing through the node contains the variable. The method may be used on neural network computation in a deep learning training system.

Abstract: A reshaping method for metal product includes acquiring position data information of a workpiece; converting the position data information into coordinate information, and fitting the coordinate information to obtain a surface contour curve of the workpiece; comparing the surface contour curve with a standard contour curve to generate a comparison result; obtaining reshaping information of the workpiece based on the comparison result; and controlling the reshaping system to reshape the workpiece. The disclosure acquires position data information of the workpiece by measuring, converts the position data information into coordinate information, fits the coordinate information to obtain the surface contour curve of the workpiece, compares the surface contour curve with the standard contour curve to obtain deformation situation, and reshapes the workpiece according to the reshaping information based on the deformation situation, which improves efficiency of workpiece reshaping and reducing labor costs.

Abstract: A cleaning apparatus includes a device body; a bumper, connected to the device body and comprising a bumper opening, provided with a photography opening and an illuminator opening; a camera, oriented outside of the device body through the photography opening, configured to pick up an image of the environment to perform functions of object recognition; and an illuminator, oriented outside of the device body through the illuminator opening, configured to emit light to illuminate at least part of a recognition area of the camera.

Abstract: A display substrate, a manufacturing method thereof, and a display device are provided. The display substrate includes a base substrate and a white OLED display unit on the base substrate, and further includes: an optical adjustment structure on a light emitting side of the white OLED display unit, where the optical adjustment structure is in a peripheral region of each pixel region. The optical adjustment structure is configured to absorb light in a first wavelength range or convert light in a first wavelength range into light in a second wavelength range.

Abstract: The present disclosure discloses a distributed model compilation system. A master node of the system determines the logic calculation graph of the model based on model information, divides the logic calculation graph into multiple logic calculation sub-graphs, generates a distributing message for each logic calculation sub-graph, and then transmits the distributing message to a slave node. Each of the slave nodes allocates a local computing resource to compile the logic calculation sub-graph based on the received distributing message, and transmits compilation completion information to the master node. The master node determines the completion of model compilation based on the compilation completion information returned by each slave node, and executes the target work based on the compiled model.

Abstract: A method for preparing Portland cement includes: respectively weighing iron slag, copper slag, vanadium slag, and nickel slag and grinding, to yield prefabricated iron slag, prefabricated copper slag, prefabricated vanadium slag, and prefabricated nickel slag; weighing mica and kaolinite, mixing, and grinding to obtain aluminous raw materials; evenly mixing the prefabricated iron slag and the aluminous raw materials, and calcining, to yield an iron-aluminum eutectic mineral; weighing the marble, fluorite, dolomite, and quartz, evenly mixing the marble, fluorite, dolomite, and quartz with the prefabricated copper slag, prefabricated vanadium slag, and prefabricated nickel slag to yield a first mixture; grinding the iron-aluminum eutectic mineral to yield powders, and calcining a second mixture of the first mixture and the powders, to yield the cement clinker; and cooling the cement clinker, and grinding a third mixture of the cooled cement clinker and the gypsum, to yield the Portland cement.

Abstract: The disclosure discloses a graph optimization method and apparatus for neural network computation. The graph optimization method includes the following steps: S1: converting a computation graph; S2: allocating a register; S3: defining a route selector for a redefined variable; S4: solving the route selector for the redefined variable; S5: defining a criterion of inserting the route selector for the redefined variable into a node; S6: analyzing a dominating edge set of the node for the redefined variable; S7: inserting the route selector for the redefined variable; and S8: renaming the redefined variable. The disclosure solves the problem of the corresponding route selection on a correct definition of the redefined variable when a node including the redefined variable in a computation graph in the compiling period flows through multiple paths of computation flow, reduces the memory cost and promotes the development of implementation application of a deep neural network model.

Abstract: A method and a composite for evaporative cooling are provided. The method includes synthesizing MOF-801 and preparing CaCl2@MOF-801 composite based on the MOF-801. The synthesizing MOF-801 includes dissolving fumaric acid and ZrOCl2·8H2O into a solvent having N, N-Dimethylformamide and formic acid to produce a mixture; heating the mixture at a predetermined temperature for a predetermined amount of time; cooling the mixture to room temperature to obtain precipitate of MOF-801; separating the MOF-801 by a filter of a predetermined pore size; and drying the separated MOF-801 at a predetermined temperature for a predetermined amount of time to activate the MOF-801. The preparing CaCl2@MOF-801 composite includes dissolving a predetermined amount of CaCl2 in deionized (DI) water; applying ultrasonication to the solution for a predetermined amount of time; and mixing the MOF-801 synthesized with the CaCl2 solution under ultrasonication at a predetermined temperature for a predetermined amount of time.

Abstract: A nanofiber actuator comprises a composite structure and a vanadium dioxide layer. The composite structure comprises a carbon nanotube wire and an aluminum oxide layer. The aluminum oxide layer is coated on a surface of the carbon nanotube wire, and the aluminum oxide layer and the carbon nanotube wire are located coaxially with each other. The vanadium dioxide layer is coated on a surface of the composite structure, and the vanadium dioxide layer and the composite structure are located non-coaxially with each other.

Abstract: A method for adjusting brightness of a display panel and a device for adjusting brightness of a display panel are provided. The display panel includes a plurality of sub-pixels, and the method includes adjusting data voltages of at least two sub-pixels of the plurality of sub-pixels to be different when performing a zero grayscale display.

Abstract: The combination of: a portable electronic device; a base on the portable electronic device and having at least one connector part; and an accessory usable with the portable electronic device and having at least one connector part. The at least one connector part on the base and at least one connector part on the accessory are configured to cooperate such that: a) the accessory is movable from a starting position, fully spaced from the base, to against the base and guidingly relative to the base in translation in a first direction along a first path portion towards one fully connected relationship with the base; and b) with the accessory in the one fully connected relationship with the base the accessory is blocked from being separated from the base.