Abstract: The present disclosure provides a driving module, a driving method, and a display device. The driving module includes multiple stages of driving circuits; the driving circuit includes an input circuit, a first reset circuit, and a second reset circuit. The input circuit controls the potential of the pull-up node based on the input signal. The first reset circuit, under the control of the first reset signal, inputs a first voltage signal to the pull-up node. The first pull-down noise reduction circuit, under the control of the potential of the first pull-down node, inputs a second voltage signal to the pull-up node. The voltage value of the first voltage signal is greater than the voltage value of the second voltage signal.

Abstract: Method of optimizing well placement in an unconventional reservoir, by obtaining a plurality of produced oil, produced water and produced gas samples from an unconventional reservoir over a period of time, and obtaining a plurality of rock samples from the reservoir. Each of those plurality of samples is chemically fingerprinted, as well as assigned time and location identifiers. This data is then used to generate a plurality of reservoir maps over time and those maps then used to optimize well placement in the reservoir.

Abstract: Methods, systems, and apparatuses are provided to track features across multiple images for use in various systems. For example, a computing device receives at least a first image and a second image captured by a camera, and detects a feature within each of the first image and the second image. The feature is located at a first feature position within the first image and at a second feature position within the second image. The computing device also receives a first sensor pose of the sensor used to capture the first image and a second sensor pose of the sensor used to capture the second image. The computing device determines a portion of third image based on the first sensor pose, the second sensor pose, the first feature position, and the second feature position. The computing device then generates feature detection data characterizing whether the feature is detected.

Abstract: An array substrate, a display panel and a manufacturing method thereof are provided. The array substrate includes data lines, gate lines, a gate driving structure electrically connected with the gate lines and the gate driving signal lines, gate driving signal lines and a data fanout wiring region. The array substrate further includes a dummy fanout wiring region, a dummy pad region and a bended line, the data fanout wiring region and the dummy fanout wiring region are located at both sides of the display region, the data lines are electrically connected with pads in the dummy pad region through the wires in the dummy fanout wiring region; the gate driving signal lines are electrically connected with pads in the dummy pad region; a transmission line is disposed at one side of the gate driving structure away from the display region and electrically connected with the bended line.

Abstract: Provided is a method for autonomous unmanned aerial vehicle (UAV) landing, including: collecting an unmanned vehicle image dataset in advance, training an unmanned vehicle detection model by using the unmanned vehicle image dataset combined with a YOLOv5 neural network; collecting, by the UAV during a landing process, images of an area below the UAV at specified time intervals, inputting the collected images into the unmanned vehicle detection model for recognition and detection; if an unmanned vehicle is recognized, further determining position information of the unmanned vehicle, and outputting, by a control module, a long-range guidance control instruction, to instruct the UAV to fly to a specified distance position above the unmanned vehicle; collecting, by the UAV, an image of a target and determining position information of the target, and outputting, by the control module, a short-range guidance control instruction to instruct the UAV to land on an unmanned vehicle platform.

Abstract: An array substrate, a method for manufacturing the array substrate, and a display apparatus; the array substrate includes a base substrate, including multiple sub-pixel areas arranged in an array; multiple data lines, disposed at column gaps of the multiple sub-pixel areas, where at least part of the data lines include a widened portion for supporting a photo spacer; and multiple gate lines and multiple common electrode lines, disposed at row gaps of the plurality of sub-pixel areas; where the multiple gate lines and the multiple common electrode lines are disposed on a different layer from the multiple data lines, and one of part of the multiple gate lines and at least part of the multiple common electrode lines includes an avoidance portion wrapped around the widened portion.

Abstract: Various embodiments are directed to configuring or training deep neural network (DNN) machine learning models comprising one or more hidden layers and an output layer. Various embodiments provide technical advantages in training DNN machine learning models, including improved computational efficiency and guaranteed optimality. In one embodiment, an example method includes identifying a nonlinear-model-based representation for each hidden layer, which may be a Bank of Wiener Models, a nonlinear units of the hidden layer, and/or the like. The method further includes individually and sequentially configuring the hidden layers, each configured by determining a correlation measure (e.g., a correlation ratio) between the layer output and a target signal. Parameters of the particular hidden layer are modified by maximizing the correlation measure to yield maximal correlation over the space of functions.

Abstract: A convolution circuit includes a plurality of multipliers, a first adder coupled to the plurality of multipliers, and a second adder coupled to the first adder. Each multiplier includes a plurality of precoders, a plurality of encoder groups, and an adder tree circuit. Each precoder is in a one-to-one correspondence with one encoder group. Output ends of the plurality of encoder groups and input lines of the adder tree circuit are of a same quantity and in a one-to-one correspondence. In addition, the adder tree circuit is coupled to the first adder. The second adder is further coupled to a memory. A partial product that is related only to a weight parameter may be first accumulated with a constant 1 in the multiplier, and then added to results output by adder tree circuits in the second adder.

Abstract: A method includes segmenting a source video into source video segments, and generating a script for a new video using a generative artificial intelligence (AI) engine. The script includes, for each of one or more new video segments arranged according to a sequential order, a segment descriptor and a segment voice-over transcript. For each new video segment, a voice-over segment is generated from among the source video segments based on the respective segment voice-over transcript, and a set of source video segment(s) is selected based on the respective segment descriptor, for use in generating the new video segment. The method also includes generating the new video, at least in part by inserting the generated voice-over segments for the new video segment(s), and the selected set(s) of source video segment(s) for the new video segment(s), in accordance with the sequential order.

Abstract: A method for generating a new video from a source video includes determining that the source video is associated with one or more components, and identifying a source starting segment within the source video at least in part by selecting a segment identification model, from among a plurality of candidate segment identification models, based at least in part on the segment identification module being configured to operate upon at least one of the one or more components. The method also includes identifying the source starting segment by using the selected segment identification model to process at least a portion of the source video. The method also includes generating the new video using one or more portions of the source video, wherein generating the new video includes generating an initial segment of the new video based on the source starting segment.

Abstract: Disclosed are methods and compositions for treating cell proliferative diseases and disorders such as cancers. Particularly disclosed are methods and composition for treating cancers such as glioblastoma by administering a therapeutic agent that inhibits the biological activity of the autophagy related 4B cysteine peptidase (ATG4B) protein in conjunction with additional therapeutic agents or treatments.

Abstract: Systems, methods, and apparatus, including computer-readable media, for managing and deploying customized artificial intelligence chatbots. In some implementations, a system receives data indicating user input instructing a chatbot to be created and an indication of a data set for the chatbot to use. The system creates the chatbot based on data objects in the data set. The system provides a code segment or module configured to cause a chatbot interface for interacting with the chatbot to be embedded in a user interface. The system receives a user prompt provided for the chatbot through the user interface. The system provides a response to the user prompt from the chatbot, and the response from the chatbot includes text generated by one or more artificial intelligence and/or machine learning (AI/ML) models using values from the data set.

Abstract: Implementations claimed and described herein provide systems and methods for managing natural resource production. The systems and methods use a machine learning model to generate categorizations associated with communication data. The machine learning model is built from historical data.

Abstract: A star rating management method for deployment and implementation of gas field development wells first establishes the expression for estimating the gas well initial production rate and the dimensionless production curve; then determines the annual production of the new well for every year during the evaluation period and determines the internal rate of return of the new gas well. If the internal rate of return fails to meet the requirement, the deployment location and construction technology of the new well are re-optimized. Corresponding stars rating are determined according to the internal rate of return of the gas wells. The new well is drilled according to the deployment location that can meet the economic benefit requirements. After the gas well is drilled, a more detailed fracturing design is carried out, and the internal rate of return of the gas well and the star rating of development benefit are redetermined.

Abstract: Disclosed are a device and method for determining gas diffusion coefficient in rocks. The device includes: a core holder, a sampling mechanism, a gas analyzer, a gas source and a pressure gauge. A middle part of the core holder holds a rock sample, two sampling mechanisms are symmetrically arranged at two ends of the rock sample, each sampling mechanism includes a first blind tube, a cylinder, a second blind tube and a sampling cavity, an open end of the first blind tube faces the rock sample and is in sealing connection with an end surface of the rock sample to form a diffusion chamber. A single sampling cavity is isolated from the diffusion chamber, and the pressure of the diffusion chamber is not affected, so that undisturbed sampling is achieved, the influence on the pressure of the diffusion chamber is reduced, and the accuracy of the determination results is improved.

Abstract: A model-as-a-service (MaaS) platform performs cross-model resources allocation from a shared pool of GPU resources based on model-agnostic metrics generated by a metric standardizer. The metric standardizer receives, from model providers, model-specific benchmark metrics that define relationships between resource utilization and token processing according to the different model-specific tokenization schemes; receives, from one or more MaaS components, token-based job metrics pertaining to LLM processing tasks; and determines, based on the model-specific benchmark metrics and token-based job metrics, the model-agnostic metrics for multiple model pools executing instances of different large language models (LLMs) that generate and process text according to different model-specific tokenization schemes. The MaaS platform further includes one or more resource allocation components that dynamically reallocates resources of the shared pool based on the model-agnostic metric.

Abstract: A model-as-a-service (MaaS) platform performs cross-model resources allocation from a shared pool of GPU resources based on model-agnostic metrics generated by a metric standardizer. The metric standardizer receives, from model providers, model-specific benchmark metrics that define relationships between resource utilization and token processing according to the different model-specific tokenization schemes; receives, from one or more MaaS components, token-based job metrics pertaining to LLM processing tasks; and determines, based on the model-specific benchmark metrics and token-based job metrics, the model-agnostic metrics for multiple model pools executing instances of different large language models (LLMs) that generate and process text according to different model-specific tokenization schemes. The MaaS platform further includes one or more resource allocation components that dynamically reallocates resources of the shared pool based on the model-agnostic metric.

Abstract: A model-as-a-service (MaaS) platform performs cross-model resources allocation from a shared pool of GPU resources based on model-agnostic metrics generated by a metric standardizer. The metric standardizer receives, from model providers, model-specific benchmark metrics that define relationships between resource utilization and token processing according to the different model-specific tokenization schemes; receives, from one or more MaaS components, token-based job metrics pertaining to LLM processing tasks; and determines, based on the model-specific benchmark metrics and token-based job metrics, the model-agnostic metrics for multiple model pools executing instances of different large language models (LLMs) that generate and process text according to different model-specific tokenization schemes. The MaaS platform further includes one or more resource allocation components that dynamically reallocates resources of the shared pool based on the model-agnostic metric.

Abstract: An array substrate includes a driving circuit arranged on the base substrate, and the driving circuit includes a pull-up node control circuit, a first pull-down node control circuit, and an output circuit; the pull-up node control circuit controls a potential of the pull-up node; the first pull-down node control circuit controls to write a first control voltage provided by the first control voltage line into the first pull-down node; the output circuit controls the driving signal output terminal to output a driving signal under the control of the potential of the pull-up node; the array substrate also includes a first conductive portion arranged on the base substrate; the first conductive portion is electrically connected to the first control voltage line.