Abstract: A medical image detection system includes a memory and a processor. The processor is configured to execute the neural network model stored in the memory. The neural network model includes a feature extractor, a feature pyramid network, a first output head and a second output head. The feature extractor is configured for extracting intermediate tensors from a medical image. The feature pyramid network is associated with the feature extractor. The feature pyramid network is configured for generating multi-resolution feature maps according to the intermediate tensors. The first output head is configured for generating a global prediction according to the multi-resolution feature maps. The second output head is configured for generating local predictions according to the multi-resolution feature maps. The processor is configured to generate output information based on the medical image, the global prediction and the local predictions.

Abstract: A fan assembly for use in a sidewall of an animal house. The fan assembly includes an impeller, a motive force configured to rotate the impeller about a fan assembly axis, and a butterfly shutter. A shutter operation mechanism moves the butterfly doors of the shutter between the closed position and the open position in synchrony with an operating speed of the impeller. The shutter operating mechanism includes pushrods connected to the butterfly doors, and a linkage between the impeller and the second pushrods. The linkage includes at least one weight configured to rotate with the impeller about the fan assembly axis, wherein rotation of the at least one weight causes the linkage to exert a force on the pushrods to push the butterfly doors toward the open position when the impeller is rotating.

Abstract: A control method for a reinforcement learning system includes following operations. The reinforcement learning system obtains training data relating to an interaction system. The interaction system interacts with a reinforcement learning agent. A neural network model is utilized by the reinforcement learning agent for selecting sequential actions from a set of candidate actions. The neural network model is trained to maximize cumulative rewards collected by the reinforcement learning agent in response to the sequential actions. During training of the neural network model, auxiliary rewards of the cumulative rewards are provided to the reinforcement learning agent according to a comparison between symptom inquiry actions of the sequential actions and diagnosed symptoms in the training data.

Abstract: A method of forming a semiconductor device includes forming a first layer over a substrate in a deposition chamber with a first deposition cycle and forming a second layer over the substrate in the deposition chamber with a second deposition cycle. The first deposition cycle includes flowing a first process gas over the substrate and flowing a second process gas over the substrate. The second deposition cycle includes flowing a third process gas over the substrate and flowing a fourth process gas over the substrate.

Abstract: The present application is related to system and method for analyzing confidential data. Firstly, a first key is used to obtain a first analysis authorization for proceeding a first analysis responsive to an operational model in an encrypted cloud space with a connection of a network. Then, the result of the first analysis is verified. While the verifying of the operational model is pass, a second key is used to obtain a second analysis authorization for proceeding a second analysis responsive to an operational model without the connection of the network. Thereby, the cloud technique for analyzing data can be applied for analyzing confidential data.

Abstract: A medical system is able to provide a symptom query interpretation and/or a disease diagnosis interpretation. The medical system includes an interface and a processor. The interface is configured for receiving an input state. The processor is coupled with the interface. The processor is configured to execute a symptom checker to select a current action, from a plurality of candidate symptom queries and a plurality of candidate disease predictions, according to the input state. In response to the current action is a first symptom query, the processor is configured to execute an interpretable module interacted with the symptom checker to generate a diagnostic tree for simulating possible diagnosis paths, and generate a symptom query interpretation about the first symptom query according to the diagnostic tree.

Abstract: Provided are functional segregated telodendrimers having, for example, two or three functional segments. The telodendrimers can have one or more crosslinking groups (e.g., reversible photocrosslinking groups). The telodendrimers can aggregate to form nanocarriers. Cargo such as drugs, imaging probes, and other materials may be sequestered in the core of the aggregates via non-covalent or covalent interactions with the telodendrimers. Such nanocarriers may be used in drug delivery applications and imaging applications.

Abstract: A layout arrangement of a driver integrated circuit includes multiple output pads, a plurality of switching circuits, and multiple data channel circuits. The output pads include a first output pad and a second output pad and are configurable to be coupled to a plurality of data lines. The switching circuits include a first switching circuit. A first selection terminal of the first switching circuit is coupled to the first output pad via a first connecting wire. A second selection terminal of the first switching circuit is coupled to the second output pad via a second connecting wire. The data channel circuits include a first data channel circuit. An output terminal of the first data channel circuit is coupled to a common terminal of the first switching circuit via a third connecting wire. The third connecting wire is longer than the first connecting wire and the second connecting wire.

Abstract: A driver integrated circuit, including multiple output pads, multiple switching circuits, and multiple data channel circuits, is provided. The output pads and the switching circuits are arranged in a pad area of the driver integrated circuit. The output pads include a first output pad and a second output pad, and the switching circuits include a first switching circuit. A first selection terminal of the first switching circuit is coupled to the first output pad. A second selection terminal of the first switching circuit is coupled to the second output pad. The data channel circuits are arranged in a function circuit area of the driver integrated circuit. The data channel circuits include a first data channel circuit. An output terminal of the first data channel circuit is coupled to a common terminal of the first switching circuit.

Abstract: A machine learning method includes steps of: obtaining, by a processor, a model parameter from a memory, and performing, by a processor, a classification model according to the model parameter, wherein the classification model comprises a plurality of neural network structural layers; calculating, by the processor, a first loss and a second loss according to a plurality of training samples, wherein the first loss corresponds to an output layer of the plurality of neural network structural layers, and the second loss corresponds to one, which is before the output layer, of the plurality of neural network structural layers; and performing, by the processor, a plurality of updating operations for the model parameter according to the first loss and the second loss to train the classification model.

Abstract: The present application is related to system and method for analyzing confidential data. Firstly, a first key is used to obtain a first analysis authorization for proceeding a first analysis responsive to an operational model in an encrypted cloud space with a connection of a network. Then, the result of the first analysis is verified. While the verifying of the operational model is pass, a second key is used to obtain a second analysis authorization for proceeding a second analysis responsive to an operational model without the connection of the network. Thereby, the cloud technique for analyzing data can be applied for analyzing confidential data.

Abstract: A training method suitable for a reinforcement learning system with a reward function to train a reinforcement learning model and including: defining at least one reward condition of the reward function; determining at least one reward value range corresponding to the at least one reward condition; searching for at least one reward value from the at least one reward value range by a hyperparameter tuning algorithm; and training the reinforcement learning model according to the at least one reward value.

Abstract: A semiconductor device includes a substrate, a first polysilicon structure over a first portion of the substrate, and a first spacer on a sidewall of the first polysilicon structure. The first spacer has a concave corner region between an upper portion and a lower portion. The semiconductor device includes a second polysilicon structure over a second portion of the substrate. The semiconductor device includes a second spacer on a sidewall of the second polysilicon structure. The semiconductor device further includes a protective layer covering an entirety of the first spacer and the first polysilicon structure, wherein the protective layer has a first thickness over the concave corner region and a second thickness over the first polysilicon structure, a difference between the first thickness and the second thickness is at most 10% of the second thickness, and the protective layer exposes a top-most portion of a sidewall of the second spacer.

Abstract: A ventilation shutter has a frame providing an opening between the exterior and interior of a building. The shutter has a plurality of louvers pivotally mounted in the frame such that each louver is moveable between an open position and a closed position where the bottom edge of one louver engages the top edge of the next lower louver. The top edge of each louver has a sealing lip extending therefrom configured to engage with the bottom edge of the adjacent louver when the plurality of louvers are in the closed position.

Abstract: An extrusion forming apparatus includes a mold disposed in a front open end of an extruding channel of a housing, a holding seat disposed in front of the front open end and partially extending into the extruding channel to abut the mold, and a vibrating source connected to the housing in abutment with the holding seat and controllable to apply an ultrasonic energy to the mold through the holding seat. An extrusion forming method is carried out by the extrusion forming apparatus. An extruded product is made by the extrusion forming method using the extrusion forming apparatus.

Abstract: A fan assembly for use in a sidewall of an animal house. The fan assembly includes an impeller, a motive force configured to rotate the impeller about a fan assembly axis, and a butterfly shutter. A shutter operation mechanism moves the butterfly doors of the shutter between the closed position and the open position in synchrony with an operating speed of the impeller. The shutter operating mechanism includes pushrods connected to the butterfly doors, and a linkage between the impeller and the second pushrods. The linkage includes at least one weight configured to rotate with the impeller about the fan assembly axis, wherein rotation of the at least one weight causes the linkage to exert a force on the pushrods to push the butterfly doors toward the open position when the impeller is rotating.

Abstract: Control methods of a channel setting module applied to a display panel are provided. The display panel has gate lines, source lines, and pixels. The pixels are arranged in matrix. The pixels disposed at the same row are electrically connected to the same gate line, and the pixels disposed at the same column are electrically connected to the same source line. The adoption of the channel setting module reduces the control signals required by the source lines. The channel setting module includes operational amplifiers and de-mux switches, and the control methods dynamically determine conduction states of the de-mux switches. The voltage outputs of the operational amplifiers are selectively outputted to the source lines, depending on conduction statuses of the de-mux switches. By applying the control methods, the interference between the source lines are reduced, and the instantaneous overshoots/undershoots of floating channels are depressed.

Abstract: A manufacture includes a substrate comprising a first portion and a second portion. The manufacture further includes a first polysilicon structure over the first portion of the substrate. The manufacture further includes a second polysilicon structure over the second portion of the substrate. The manufacture further includes two spacers on opposite sidewalls of the second polysilicon structure, wherein each spacer of the two spacers has a concave corner region between an upper portion and a lower portion. The manufacture further includes a protective layer covering the first portion of the substrate and the first polysilicon structure, the protective layer exposing the second portion of the substrate, the second polysilicon structure, and partially exposing the two spacers.

Abstract: A data driver for driving a display panel includes a first driving channel coupled to a polarity inversion circuit and configured to generate a first data voltage signal having a positive polarity output to the display panel according to a plurality of first pixel data; a second driving channel coupled to the polarity inversion circuit and configured to generate a second data voltage signal having a negative polarity output to the display panel according to a plurality of second pixel data; wherein the first data voltage signal is output to first output node through the polarity inversion circuit during a first line period and the second data voltage signal is output to the first output node through the polarity inversion circuit during a second line period after the first line period, and the first line period and the second line period respectively belong to two consecutive frame periods.

Abstract: A driving integrated circuit (IC) and a fan-out compensation method thereof are provided. The driving IC includes a plurality of driving channel circuits, a plurality of output buffer circuits and a compensation control circuit. The input terminals of the output buffer circuits are coupled to the output terminals of the driving channel circuits in a one-to-one manner. The output terminals of the output buffer circuits are coupled a plurality of data lines of a display panel in a one-to-one manner. The compensation control circuit is coupled to the output buffer circuits for adjusting the slew rate of the output terminals of the output buffer circuits to compensate difference in delay times between the data lines of the display panel.