Abstract: A virtual machine backup method, performed by a first host, includes: capturing a request to write data from a virtual machine to a hard disk image file, wherein the request includes written data and input and output location information, copying the written data to a temporary storage area, calculating a first key of the written data, storing the first key, the input and output location information into a first resource location structure, pausing an operation of the virtual machine and generating a second resource location structure according to the first resource location structure, the first key and a second key, and outputting a backup data set to a second host according to the second resource location structure, wherein the backup data set includes the second resource location structure and only one of existing data and the written data when the first key and the second key are the same.

Abstract: A light-transmitting antenna includes a substrate, a first and a second conductive pattern. The first and the second conductive pattern is disposed on a first and a second surface of the substrate respectively. The first conductive pattern includes a first feeder unit, a first and a second radiation unit, a first and a second coupling unit and a first parasitic unit. The first feeder unit is connected to the second radiation unit. The first and the second radiation unit are located between the first and the second coupling unit. One side and the other side of the first parasitic unit is connected to the second coupling unit and adjacent to the first coupling unit respectively. The second conductive pattern includes a second feeder unit, a third coupling unit, a second parasitic unit, and a fourth coupling unit.

Abstract: Disclosed is a method for enhancing memory utilization and throughput of a computing platform in training a deep neural network (DNN). The critical features of the method includes: calculating a memory size for every operation in a computational graph, storing the operations in the computational graph in multiple groups with the operations in each group being executable in parallel and a total memory size less than a memory threshold of a computational device, sequentially selecting a group and updating a prefetched group buffer, and simultaneously executing the group and prefetching data for a group in the prefetched group buffer to the corresponding computational device when the prefetched group buffer is update. Because of group execution and data prefetch, the memory utilization is optimized and the throughput is significantly increased to eliminate issues of out-of-memory and thrashing.

Abstract: An air-floating guide rail device includes a guide rail unit, a slider unit, and a linear motor unit. The guide rail unit includes a guide rail body and two air-floating block sets made of a material different from that of the guide rail body and each including top and side air-floating blocks. The slider unit includes a main sliding seat and two lateral sliding seats connected integrally to the main sliding seat and each having first and second guiding surfaces transverse to each other and disposed respectively adjacent to corresponding top and side air-floating blocks, and first and second air guiding passages connecting the first and second guiding surfaces to the external environment. The linear motor unit includes a stator and a mover mounted fixedly to the main sliding seat and movable relative to the stator for driving linear movement of the slider unit relative to the guide rail unit.

Abstract: A light-transmitting antenna includes a substrate, a first conductive pattern, and a second conductive pattern. The first conductive pattern has a first feeder unit, a first radiation unit, a second radiation unit, and a first connection unit. The first feeder unit and the first connection unit are connected to two sides of the first radiation unit. The first connection unit connects the first radiation unit and the second radiation unit. The second conductive pattern has a second feeder unit, a third radiation unit, a fourth radiation unit, and a second connection unit. The second feeder unit and the second connection unit are connected to two sides of the third radiation unit. The second connection unit connects the third radiation unit and the fourth radiation unit. An orthogonal projection of the second feeder unit on a first surface of the substrate at least partially overlaps the first feeder unit.

Abstract: A device for estimating breast implant volume is provided, which may include an input interface and a processing circuit. A numerical value may be inputted in to the input interface. The processing circuit may calculate the estimation value of breast implant volume according to a linear model and the numerical value. The processing circuit may multiply the numerical value by a coefficient to generate a product, and add a constant to the product to generate the estimation value of breast implant volume, wherein the numerical value may be a breast sample weight or a breast size.

Abstract: An impeller is provided, including a metal housing, a shaft, and a plastic member. The metal housing has a shaft mounting hole. The inner surface of the shaft mounting hole includes three or more contact points, and the contact points are closer to the shaft than other portions of the inner surface of the shaft mounting hole. The shaft passes through the shaft mounting hole and is affixed by the contact points. The metal housing divides the shaft into an upper section, a middle section, and a lower section. The plastic member passes through the shaft mounting hole and is in contact with the middle section.

Abstract: A radiation source apparatus includes a vessel, a laser source, a collector, a horizontal obscuration bar, and a reflective mirror. The vessel has an exit aperture. The laser source is configured to emit a laser beam to excite a target material to form a plasma. The collector is disposed in the vessel and configured to collect a radiation emitted by the plasma and to reflect the collected radiation to the exit aperture of the vessel. The horizontal obscuration bar extends from a sidewall of the vessel at least to a position between the laser source and the exit aperture of the vessel. The reflective mirror is in the vessel and connected to the horizontal obscuration bar.

Abstract: Methods for reducing contact resistance include performing a selective titanium silicide (TiSi) deposition process on a middle of the line (MOL) contact structure that includes a cavity in a substrate of dielectric material. The contact structure also includes a silicon-based connection portion at a bottom of the cavity. The selective TiSi deposition process is selective to silicon-based material over dielectric material. The methods also include performing a selective deposition process of a metal material on the MOL contact structure. The selective deposition process is selective to TiSi material over dielectric material and forms a silicide capping layer on the silicon-based connection portion. The methods further include performing a seed layer deposition process of the metal material on the contact structure.

Abstract: Disclosed is a method for enhancing memory utilization and throughput of a computing platform in training a deep neural network (DNN). The critical features of the method includes: calculating a memory size for every operation in a computational graph, storing the operations in the computational graph in multiple groups with the operations in each group being executable in parallel and a total memory size less than a memory threshold of a computational device, sequentially selecting a group and updating a prefetched group buffer, and simultaneously executing the group and prefetching data for a group in the prefetched group buffer to the corresponding computational device when the prefetched group buffer is update. Because of group execution and data prefetch, the memory utilization is optimized and the throughput is significantly increased to eliminate issues of out-of-memory and thrashing.

Abstract: Disclosed are support-activators and catalyst compositions comprising the support-activators for polymerizing olefins in which the support-activator includes clay heteroadduct, prepare from a colloidal phyllosilicate such as a colloidal smectite clay, which is chemically-modified with a heterocoagulation agent. By limiting the amount of heterocoagulation reagent relative to the colloidal smectite clay as described herein, the smectite heteroadduct support-activator is a porous and amorphous solid which can be readily isolated from the resulting slurry by a conventional filtration process, and which can activate metallocenes and related catalysts toward olefin polymerization. Related compositions and processes are disclosed.

Abstract: A system for obstacle detection adapted to a self-guiding machine is provided. The system includes a controller, a linear light source and a light sensor. The linear light source and the light sensor are set apart at a distance. When the linear light source emits an indicator light being a vertical linear light projected onto a path the self-guiding machine travels toward, the light sensor senses the indicator light. The vertical linear light is segmented into a first segment projected to a ground and a second segment projected to a floating obstacle when the self-guiding machine approaches the floating obstacle with a height from the ground and the indicator light is projected to the floating obstacle, in which the second segment of the light sensed by the light sensor is determined as the floating obstacle in front of the self-guiding machine.

Abstract: A data writing method, a memory storage device, and a memory control circuit unit are provided. The method includes: receiving a write command from a host system, and the write command including first data; checking a status of a first physical programming unit in a first physical erasing unit; in response to the status of the first physical programming unit being a first status, sending a first command sequence to a rewritable non-volatile memory module, and the first command sequence being configured to instruct the rewritable non-volatile memory module to store at least part of the first data to the first physical programming unit.

Abstract: A method for manufacturing a semiconductor device includes: providing a wafer-bonding stack structure having a sidewall layer and an exposed first component layer; forming a photoresist layer on the first component layer; performing an edge trimming process to at least remove the sidewall layer; and removing the photoresist layer. In this way, contaminant particles generated from the blade during the edge trimming process may fall on the photoresist layer but not fall on the first component layer, so as to protect the first component layer from being contaminated.

Abstract: A method for manufacturing a semiconductor structure is provided. The method includes: providing a substrate and a dielectric layer on the substrate; forming a hole in the dielectric layer; forming an initial barrier material layer and a conductive layer on an upper surface of the dielectric layer and in the hole; removing part of the initial barrier material layer and part of the conductive layer to form a barrier material layer and a via element in the hole respectively and expose the upper surface of the dielectric layer. An upper surface of the barrier material layer is higher than the upper surface of the dielectric layer.

Abstract: An electronic device including a pixel array structure, a redistribution structure, and a plurality of conductive via structures is provided. The pixel array structure includes a plurality of signal lines. The redistribution structure overlaps the pixel array structure and includes a plurality of conductive lines. The conductive via structures electrically connect the signal lines of the pixel array structure and the conductive lines of the redistribution structure. At least one of the conductive via structures shares at least one conductive layer with the pixel array structure.

Abstract: This disclosure relates to a transparent display that includes a transparent substrate, a plurality of wires, and a plurality of electronic components. The wires are disposed on the transparent substrate, and the wires have at least one serpentine part having a first end and a second end that are opposite to each other. There is an extending path formed from the first end to the second end. The electronic components are disposed on the plurality of wires. The wires have at least one part that surrounds an opening area and has a total extending length. The at least one part of the wires includes the at least one serpentine part. A ratio of a sum in length of the at least one serpentine part along the extending path to the total extending length is equal to or greater than 10%.

Abstract: A backlight driving method includes steps of: (A) receiving a piece of image data that includes a number (K) of segments, where K?2; (B) generating a piece of adjustment data that includes a number (K) of segments; each segment of the adjustment data being generated based on a respective segment of the image data and upon receipt of the respective segment of the image data; (C) generating, based on a piece of delay data and on an original synchronization control (SC) signal that has a pulse, an internal SC signal that has a number (K) of pulses; and (D) generating a backlight driving output based on the adjustment data and the internal SC signal, so as to drive a backlight source of a scan-type display to emit light.

Abstract: A leak detection sensor is disclosed for detecting a leaking liquid spilled from an adapter-coupling position, and includes a first conductive layer, an insulator and a second conductive layer. The first conductive layer includes a first through hole. The adapter-coupling position is located adjacent to the first through hole. The insulator includes a second through hole. The second conductive layer is connected to the first conductive layer through the insulator, and includes a conducting surface in fluid communication with the adapter-coupling position through the second through hole and the first through hole. The first conductive layer and the second conductive layer are insulated from each other through the insulator. When the leaking liquid is spilled into the first through hole and the second through hole, and contacts the conducting surface, the second conductive layer is conducted to the first conductive layer to form a conducting-resistance value.

Abstract: A backlight driving method includes steps of: (A) generating an original synchronization control (SC) signal, and a serial input signal that contains multiple predetermined delay values; (B) generating multiple internal SC signals based on the original SC signal and the delay values, such that respective time delays of the internal SC signals with respect to the original SC signal are respectively dependent on the delay values; and (C) generating multiple backlight driving outputs based on the internal SC signals to respectively drive multiple backlight sources, such that the backlight sources emit light in an order dependent on the delay values.