Abstract: Embodiments described herein provide a large language model (LLM) based AI agent that adopts Monte-Carlo Tree Search (MCTS) to execute a task. The LLM is prompted with a task description and it responds with its first attempted list of actions. Based on the success or failure of the first attempt, the LLM is prompted with an updated prompt which includes feedback from the first attempt based on a determined reward. The prompt may include a relative “score” for each action taken at each step. A numeric score may be mapped to a set of pre-defined text labels, such as “high” or “low” value putting the score in a form more suited for an LLM prompt. In this way, the LLM is iteratively given prompts which are updated with the scores from each action taken at each previous iterations so that it traverses different paths on the tree in each iteration.

Abstract: This disclosure provides compounds containing 7-(pyrimidin-4-yl)quinolin-4(1H)-one structure, the use thereof for selectively inhibiting the activity of CDK4, and pharmaceutical compositions comprising the compounds as treatment of various diseases including cancer.

Abstract: Disclosed is a method for fast automatic calibration of a phased array based on a residual neural network. A phase setting matrix is set and an amplitude and a phase of a array far-field complex signal are measured with a network analyzer to obtain an amplitude and phase vector of the array far-field complex signal. A real part, an imaginary part, and a magnitude of the far-field measured complex signal value are separated and normalized, and mapped to RGB three-channel image data. Datasets are automatically generated according to a preset amplitude-phase error range by a simulation software, the datasets are proportionally divided into a training set and a test set to be input into the residual neural network for training to obtain a calibration model. Measured data is input into the calibration model for automatic estimation of the amplitude-phase error of the phased array.

Abstract: Provided are a display substrate and a preparation method thereof, a display panel and a mask. The display substrate includes an underlayer and an organic structure disposed on the underlayer. The organic structure includes a first surface arranged away from the underlayer; in a direction perpendicular to a plane where the underlayer is located, a difference between a maximum distance between the first surface and the underlayer and a minimum distance between the first surface and the underlayer is less than or equal to 0.06 ?m.

Abstract: The present application provides a separator, a method for preparing the same, a secondary battery and an electrical device related thereto. The separator includes a porous substrate and a coating layer disposed on one or more surfaces of the porous substrate, wherein the coating layer includes nanocellulose, and the porous substrate has a surface tension of ?1 mN/m, the coating layer has a surface tension of ?2 mN/m, and the separator satisfies ?1/?2?0.68. The separator provided in the present application has the characteristics of excellent heat resistance and high bonding strength, thus the secondary battery using the separator can have the combined characteristics of high energy density, high thermal safety performance, and long service life.

Abstract: Disclosed is a method for fast automatic calibration of a phased array based on a residual neural network. A phase setting matrix is set and an amplitude and a phase of a array far-field complex signal are measured with a network analyzer to obtain an amplitude and phase vector of the array far-field complex signal. A real part, an imaginary part, and a magnitude of the far-field measured complex signal value are separated and normalized, and mapped to RGB three-channel image data. Datasets are automatically generated according to a preset amplitude-phase error range by a simulation software, the datasets are proportionally divided into a training set and a test set to be input into the residual neural network for training to obtain a calibration model. Measured data is input into the calibration model for automatic estimation of the amplitude-phase error of the phased array.

Abstract: A speaker assembly including a first speaker comprising a first diaphragm and a first voice coil movably coupled to a first magnet assembly, wherein the first diaphragm faces a first direction, the first voice coil moves along a first axis in the first direction when driven by an audio signal and the first magnet assembly is coupled to a fixed structure by a first compliant mounting member; and a second speaker laterally offset from the first speaker and including a second diaphragm and a second voice coil movably coupled to a second magnet assembly, wherein the second diaphragm faces a second direction different from the first direction, the second voice coil moves along a second axis in the second direction when driven by an audio signal and the second magnet assembly is coupled to the fixed structure by a second compliant mounting member.

Abstract: A method for adding animation to a video includes: detecting whether the hand of a user in the current video frame, which is currently captured, holds an object; if so, determining a target animation corresponding to the object; and determining an animation addition area, which corresponds to the target animation, on the current video frame, and adding the target animation to the animation addition area. Therefore, a corresponding animation special effect is added for an object held by a user in a video, such that the interaction between the user and the handheld object is visually reflected, thereby improving the interestingness of video filming, and providing technical support for improving the viscosity of the user with respect to a video product.

Abstract: Disclosed are a vapor chamber, a radiator and an electronic device. The vapor chamber includes: a bottom plate, a top plate, a side plate and a backflow piece. The bottom plate includes a first heat dissipation area and a second heat dissipation area provided at a peripheral side of the first heat dissipation area. The first heat dissipation area is provided with at least a first capillary structure layer on the side near the top plate, and the second heat dissipation area is provided with at least a second capillary structure layer on the side near the top plate. The capillary force of the first capillary structure layer is greater than that of the second capillary structure layer, and the flow resistance of the fluid medium in the second capillary structure layer is smaller than that in the first capillary structure layer.

Abstract: Embodiments of the present disclosure provide a method for phased array calibration based on CNN-LSTM using power measurement, comprising: establishing a phased array calibration signal model, and utilizing a program to conveniently obtain a large amount of data for training a neural network without the need for actual measurements; converting and preprocessing the generated data, and saving as a training dataset in the form of feature data and labels; establishing a CNN-LSTM network, and inputting the training data with labels into the CNN-LSTM network for training until the CNN-LSTM network converges to obtain the final calibration model; measuring the phased array to be measured to obtain feature data, obtaining a calibration result of the phased array by inputting the feature data into the calibration model obtained from the training.

Abstract: The embodiments of the present disclosure provide a visual positioning method, the method may include obtaining a positioning image collected by an imaging device; obtaining a three-dimensional (3D) point cloud map associated with an area where the imaging device is located; determining a target area associated with the positioning image from the 3D point cloud map based on the positioning image; and determining positioning information of the imaging device based on the positioning image and the target area.

Abstract: This application discloses a heat-resistant protective member and a battery. The heat-resistant protective member includes a reinforcing layer, a functional layer, and a strengthening layer that are sequentially arranged; the functional layer includes a first resin and a filler dispersed in the first resin; and the reinforcing layer and the strengthening layer each include a fiber matrix. In the foregoing manner, thermal impact of a single battery during ignition and detonation can be resisted, and structural integrity under the thermal impact can still be prevented from being broken through, so as to provide effective heat insulation and protection for a battery pack.

Abstract: An inductor current reconstruction circuit of a power converter can include: a switching current sampling circuit configured to acquire at least one of a current flowing through a main power transistor and a current flowing through a rectifier transistor to generate a switching current sampling signal; an inductor current generating circuit configured to generate a reconstruction signal representing an inductor current in one complete switching cycle; and where the reconstruction signal comprises the switching current sampling signal and a current analog signal generated according to the switching current sampling signal and an inductor voltage signal representing a voltage across an inductor in the power converter.

Abstract: Embodiments of the present application disclose a battery and an electrical device. The battery includes: a battery cell provided with a pressure relief mechanism; and a protective plate disposed opposite the pressure relief mechanism, wherein the protective plate includes at least two protective sheets, and the at least two protective sheets are two different polymer matrix composite fiber sheets. The technical solution provided by the present application can protect a box of the battery from gas flow impact and high-temperature melting caused by thermal runaway of the battery, thereby enhancing the safety performance of the battery.

Abstract: A heat-resistant protective member and a battery are disclosed. The heat-resistant protective member includes a compound layer, and the compound layer includes a fiber matrix and a resin. The resin is dispersed in pores of the fiber matrix and/or a surface of the fiber matrix, and a volume proportion of the fiber matrix in the compound layer is 50% to 75%. In this way, the thermal shock when a battery cell is ignited and exploded can be resisted, and the structural integrity can still be maintained without being broken through under the thermal shock, thereby providing effective heat insulation and protection for a battery pack.

Abstract: Embodiments described herein provide a method of predicting an action by a plurality of language model augmented agents (LAAs). In at least one embodiment, a controller receives a task instruction to be performed using an environment. The controller receives an observation of a first state from the environment. The controller selects a LAA from the plurality of LAAs based on the task instruction and the observation. The controller obtains an output from the selected LAA generated using an input combining the task instruction, the observation, and an LAA-specific prompt template. The controller determines the action based on the output. The controller causes the action to be performed on the environment thereby causing the first state of the environment to change to a second state.

Abstract: A method of information processing, an electronic device and a storage medium are provided. The method includes: receiving input information to be processed; acquiring a target answer result of the information to be processed, wherein the target answer result is generated based on a generative model and a target plugin, a cooperative mode in which the generative model and the target plugin generate the target answer result is related to an implementation requirement of a capability of the target plugin, and the target plugin is a plugin that matches with the information to be processed and is used to answer the information to be processed; and displaying the target answer result.

Abstract: The present disclosure provides an information processing method and apparatus, an electronic device, and a storage medium. The method includes: acquiring a target functional tool to be displayed and a distribution style of the target functional tool, in which different target functional tools are configured to answer questions in different vertical scenarios, and the distribution style is any one of: including a functional tool identifier and a question sample, and including a functional tool identifier and a multi-round conversation sample; and displaying the target functional tool in a target page according to the distribution style of the target functional tool.

Abstract: Embodiments described herein provide for optimizing a language model (LM) agent. In at least one embodiment, and LM agent comprises an “actor” LM and a “retrospective LM which provides reflections on attempts by the actor LM. The reflections are used to update subsequent prompts to the actor LM. Optimizing the LM agent comprises fine-tuning parameters of the retrospective LM while keeping parameters of the actor LM frozen. A gradient may be determined by a change in reward from the environment based on actions taken by the actor LM with and without a reflection of the retrospective LM. Using this gradient, parameters of the retrospective LM may be updated via backpropagation.