Abstract: A method for preparing error recovery of a memory device included in a memory system includes performing a training operation of the memory device upon power-on of the memory system, booting an operating system into a normal mode by operating the memory device using a selected operation frequency of a plurality of operation frequencies based on results of the training operation, detecting an error frequency among the plurality of operation frequencies in response to a change of the selected operation frequency of the memory device by the operating system, the error frequency being an operation frequency which causes at least one error in the memory device, and storing information regarding the detected error frequency in a first register included in a memory controller associated with the memory device.

Abstract: A computing system including a plurality of processing devices configured to execute a Mixture-of-Experts (MoE) layer included in an MoE model. The processing devices are configured to, in each of a plurality of iterations, at each of the processing devices, receive a respective plurality of input tokens. Executing the MoE layer further includes, at each of the processing devices, selecting one or more destination expert sub-models associated with the input tokens. Respective numbers k of expert sub-models selected differ across the iterations. At each of the processing devices, executing the MoE layer further includes conveying the input tokens to the one or more destination expert sub-models. Executing the MoE layer further includes generating one or more respective expert sub-model outputs at the one or more destination expert sub-models. Executing the MoE layer further includes generating and outputting an MoE layer output based on the one or more expert sub-model outputs.

Abstract: The present disclosure relates to a heteroaryl derivative and uses thereof. The heteroaryl derivative of the present disclosure exhibits excellent inhibitory activity against EGFR and/or HER2, and thus may be usefully employed as a therapeutic agent for EGFR- and/or HER2-related diseases.

Abstract: Provided are a silicon nitride film etching composition, a method of etching a silicon nitride film using the same, and a manufacturing method of a semiconductor device. Specifically, a silicon nitride film may be highly selectively etched as compared with a silicon oxide film, and when the composition is applied to an etching process at a high temperature and a semiconductor manufacturing process, not only no precipitate occurs but also anomalous growth in which the thickness of the silicon oxide film is rather increased does not occur, thereby minimizing defects and reliability reduction.

Abstract: A target protein for viral vector-based anticancer therapy, and a binding molecule or a fragment thereof which specifically binds thereto are disclosed. A conformational epitope of protein A56, and a binding molecule or a fragment thereof which specifically binds thereto are disclosed. The A56-binding molecule or a fragment thereof, which binds to a conformational epitope of A56 forms a specific structural bond with A56 and shows high affinity thereto. Administering a vaccinia virus-based oncolytic virus, A56 is expressed on the cancer cell surface in various carcinomas. The A56-binding molecule or a fragment thereof effectively targets A56 that is specifically expressed on the cancer cell surface, which enables targeted therapy for cancer cells that have survived even infection with an oncolytic virus, thereby providing effective anticancer therapy.

Abstract: A computing system is provided, including a plurality of processing devices configured to execute a Mixture-of-Experts (MoE) layer included in an MoE model. The MoE layer includes a plurality of expert sub-models that each have a respective plurality of parameter values. The MoE layer is configured to be switchable between a data parallel mode and an expert-data-model parallel mode without conveying the respective parameter values of the expert sub-models among the plurality of processing devices.

Abstract: A computing system including a plurality of processing devices configured to execute a Mixture-of-Experts (MoE) layer included in an MoE model. The processing devices are configured to execute the MoE layer at least in part by, during a first collective communication phase between the processing devices, splitting each of a plurality of first input tensors along a first dimension to obtain first output tensors. Executing the MoE layer further includes processing the first output tensors at a respective a plurality of expert sub-models to obtain a plurality of second input tensors. Executing the MoE layer further includes, during a second collective communication phase between the processing devices, receiving the second input tensors from the expert sub-models and concatenating the second input tensors along the first dimension to obtain second output tensors. Executing the MoE layer further includes outputting the second output tensors as output of the MoE layer.

Abstract: Provided is a method of operating a terminal. The method includes determining a profile item applicable to the profile view for the account based on an input received by the terminal and a coordinate indicating a position where the profile item is provided on the profile view. The method includes displaying the profile item on a screen of the terminal based on the determined profile item and the determined coordinate. The method includes receiving an input related to the profile item, and displaying a visual effect corresponding to the input on the screen.

Abstract: The present invention relates to a method for producing hollow activated carbon nanofibers for activating peroxymonosulfate used in water purification; a catalyst for water purification comprising the hollow active carbon nanofibers produced by the method; and a method for purifying contaminated water using the catalyst. The production method of the present invention can easily produce hollow activated carbon nanofibers capable of rapidly purifying contaminated water by highly efficiently activating peroxymonosulfate used for water purification.

Abstract: A mobility vehicle transfer platform includes an elevation passage provided in a structure provided with a take-off and landing site for an air mobility vehicle and having an internal space extending in the vertical direction to be connected to the take-off and landing site and to allow a ground mobility vehicle to be elevated or lowered through the internal space to be connected to or separated from the air mobility vehicle, and an elevation unit provided in the internal space of the elevation passage, having an elevation portion selectively connected to the ground mobility vehicle and configured for elevating or lowering the ground mobility vehicle in the elevation passage when the ground mobility vehicle is connected to the elevation portion.

Abstract: In an embodiment of the present invention, a proton conductive oxide fuel cell comprising an electrode substrate, a proton conductive oxide electrolyte layer positioned on the electrode substrate, a proton conductive oxide reaction prevention layer positioned on the electrolyte layer, and a proton conductive oxide air electrode layer positioned on the reaction prevention layer, wherein the reaction prevention layer is composed of ABO3-? structured perovskite proton conductive oxide, may be provided.

Abstract: Embodiments of the present disclosure relate to a system and an operating method of the system. Based on some embodiments of the disclosed technology, the system may include a random access memory structured to include memory cells to store data, a cache memory configured to cache at least part of the data, and a processor in communication with the random access memory and the cache memory to access at least part of the data from the random access memory or cache memory. The system may determine a cache hit ratio for the cache memory, and may set an operating frequency of the random access memory based on the cache hit ratio.

Abstract: A secondary battery according to an embodiment of the present disclosure includes an electrode assembly to which an electrode lead is attached; a case accommodating the electrode assembly so that a portion of the electrode lead is exposed to the outside; a sealing portion formed in the case to seal the electrode assembly; and a lead film covering a portion of an outer surface of the electrode lead and interposed between the electrode lead and the sealing portion, wherein the electrode lead includes a first region not corresponding to the lead film and exposed to the outside of the case, a second region corresponding to the lead film, and a third region not corresponding to the lead film and not exposed to the outside of the case, wherein a separation groove may be located in the third region.

Abstract: A nozzle according to an embodiment includes: a metal portion, which defines a penetration hole therein; and a ceramic portion, which surrounds the metal portion, where the metal portion includes a first region and a second region having a narrower diameter than the first region, the ceramic portion surrounds the second region, and an end of the metal portion close to a discharging end of the nozzle is spaced apart from an end of the ceramic portion close to the discharging end.

Abstract: A docking control system for a vehicle and a building and a method therefor, includes a plurality of sensors including a vehicle sensor and a gate sensor, and a controller, when a vehicle approaches a gate, which controls the plurality of sensors to set amount of movement of the vehicle, to move the gate so that a vehicle door and the gate are brought into close contact with each other, and when positional satisfaction between the vehicle door and the gate is achieved, which opens the vehicle door and the gate so that an indoor space of the vehicle and an indoor space of the building are connected each other into one space.

Abstract: Provided are a method for generating a summary and a system therefor. The method according to some embodiments may include calculating a likelihood loss for a summary model using a first text sample and a first summary sentence corresponding to the first text sample, calculating an unlikelihood loss for the summary model using a second text sample and the first summary sentence, the second text sample being a negative sample generated from the first text sample, and updating the summary model based on the likelihood loss and the unlikelihood loss.

Abstract: An antenna apparatus includes: a parabolic reflector; a dielectric support pedestal; a sub-reflector connected to an upper part of the dielectric support pedestal; and a waveguide connected to a lower part of the dielectric support pedestal, wherein the parabolic reflector has a curved surface in which a ratio of a focal length to a diameter is greater than a preset value, and at least one corrugation configured to suppress a cross polarization is formed in a region of the dielectric support pedestal.

Abstract: The present inventive concept relates to a method for removing impurities in thin film and a substrate processing apparatus. The method for removing impurities in a thin film includes the steps of: providing a substrate having a thin film formed thereon in a process chamber; supplying a first gas reacting and coupling with impurities contained in the thin film, into the process chamber; exhausting a coupled product of the impurities and the first gas by depressurizing an interior of the process chamber after stopping the supply of the first gas; curing the thin film by supplying a second gas being different from the first gas into the process chamber; and stopping the supply of the second gas and exhausting the remaining second gas from the interior of the process chamber.

Abstract: The present disclosure relates to a clip apparatus for an endoscope, the clip apparatus having a simple driving part and enabling accurate and easy clip ligation and separation. The present disclosure provides a clip apparatus for an endoscope, the clip apparatus comprising: a handle part having a moving hole formed therein; a handle slider installed so as to be slidable in the handle part through the moving hole; an inner sheath rotatably connected to the inside of the handle part and extending to the outside of the handle part; a wire connected to the handle slider and disposed inside the inner sheath; an outer sheath disposed so as to cover the inner sheath; an outer handle coupled to the outer sheath so as to move the outer sheath along the length direction of the inner sheath; and a clip coupling part connected to one end of the wire and enabling the coupling of a clip.

Abstract: An apparatus for improving turning performance of a vehicle includes: a turning characteristic detection module configured to detect a turning situation based on vehicle information obtained from at least one sensor and to calculate a required driving force to be implemented in the vehicle for a turning motion; a driving force estimation module configured to estimate a limited driving force applicable to a drive wheel; and a turning characteristic control module configured to control and apply a braking force corresponding to a difference between the required driving force and the limited driving force to a motor to inhibit a wheel slip, when the required driving force is greater than the limited driving force.