Abstract: A photovoltaic device and a method for mounting a photovoltaic device are provided. The photovoltaic device includes a color steel tile, a bonding layer, and a photovoltaic assembly. The color steel tile includes an angle relaxation portion. The bonding layer is arranged on a top wall of the angle relaxation portion. The photovoltaic assembly is located on one side of the color steel tile and connected to the color steel tile through the bonding layer. The bonding layer includes a first bonding portion and a second bonding portion arranged along a first direction. In the first direction, a ratio of a dimension of the first bonding portion to a dimension of the second bonding portion ranges from 0.1 to 0.5. The first bonding portion temporarily fix the photovoltaic assembly, while the second bonding portion mainly fix the photovoltaic assembly.

Abstract: Embodiments described herein provide systems and methods for retrieval augmented generation. A neural network based language model may be provided a question as a user input. Based on the user input, semantically diverse queries may be generated for retrieval from diverse data sources. For example, a structured data source (e.g., database or knowledge base) and unstructured data (e.g., text articles) may be used to retrieve information relevant to the user input. The retrieve information may be ranked so that the most relevant information is used by the language model in generating an answer to the question in the user input. A non-retrieval based answer generated by the language model may be utilized in some embodiments in generating the final answer.

Abstract: The present disclosure provides a multi-antenna single-station stereoscopic-arrangement ultra-short baseline lightning location method and related devices.

Abstract: A method involves forming a metal layer on a metal seed layer, the metal seed layer formed on a carrier wafer. A surface of the metal layer defines a first metal bonding layer. A second metal bonding layer is provided on a target substrate having recording head subassemblies. Mating surfaces of the first and second metal bonding layers are activated and the carrier wafer is flipped and joined with the target substrate such that the first and second metal bonding layers are bonded and the metal layer is integrated with the recording head as a near-field transducer.

Abstract: A grid-driving-circuit array is applied to a display panel, the display panel is delimited into a plurality of active areas, and each of the active areas includes multiple rows of pixel units and multiple rows of grid lines. The multiple groups of grid driving units supply grid driving signals to the plurality of active areas, and each of the groups of grid driving units includes one or more grid driving circuits. The one or more grid driving circuits are configured for supplying the grid driving signals to the grid lines within the active area corresponding to the one or more grid driving circuits. One or more multiplexers, wherein each of the multiplexers includes a plurality of frame-starting-up-signal outputting units, each of the frame-starting-up-signal outputting units of the multiplexer is configured for supplying a frame starting-up signal to one of the grid driving circuits.

Abstract: Disclosed here is a serum-free composition, comprising: a coating matrix, a basal medium, and a growth supplement, as well as methods for making and using the serum-free composition. Also disclosed are kits comprising a coating matrix, a basal medium, and a concentrated stock of a growth supplement.

Abstract: One or more aspects of the method, apparatus, and non-transitory computer readable medium include obtaining image data and computing a prediction residue value for a pixel of the image data using a prediction function. An entropy value for the pixel can then be determined based on the prediction residue value using context modeling, and progressive compressed image data for the image data can be generated based on the entropy value. The compressed image data can be used to enable collaborative image editing and other image processing tasks.

Abstract: Disclosed are various approaches for using drag-and-drop to open local files remotely. In one approach, a client device renders a user interface including a first indicium corresponding to a local file of the client device and a second indicium corresponding to a remotely executed application. A drag-and-drop action of the first indicium relative to the second indicium is detected. Responsive to the drag-and-drop action, the remotely executed application opens a copy of the local file.

Abstract: The present invention relates to novel linkers containing a 2,3-disubstituted succinic group, or 2-monosubstituted, or 2,3-disubstituted fumaric or maleic (trans (E)- or cis (Z)-butenedioic), or acetylenedicarboxyl group for conjugation of a cytotoxic agent, and/or one or more different functional molecules per linker to a cell-binding molecule, through bridge linking pairs of thiols on the cell-binding molecule specifically. The invention also relates to methods of making such linkers, and of using such linkers in making homogeneous conjugates, as well as of application of the conjugates in treatment of cancers, infections and autoimmune disorders.

Abstract: Dialogue skeleton assisted prompt transfer for dialogue summarization techniques are described that support training of a language model to perform dialogue summarization in a few-shot scenario. A processing device, for instance, receives a training dataset that includes training dialogues. The processing device then generates dialogue skeletons based on the training dialogues using one or more perturbation-based probes. The processing device trains a language model using prompt transfer between a source task, e.g., dialogue state tracking, and a target task, e.g., dialogue summarization, using the dialogue skeletons as supervision. The processing device then receives an input dialogue and uses the trained language model to generate a summary of the input dialogue.

Abstract: The disclosure provides techniques for peer to peer communication of a remote display protocol via a real-time communication session between a client device and a remote device. Embodiments include receiving, at a remote device configured to execute a virtual desktop, a request to join a real-time communication (RTC) session with a client device. Embodiments include joining the RTC session with the client device in response to receiving the request. Embodiments include receiving, at the remote device and via the RTC session, data indicative of a remote display protocol for rendering the virtual desktop on a display of the client device. Embodiments include performing, at the remote device, one or more actions based, at least in part, on receiving the data indicative of the remote display protocol.

Abstract: A method and apparatus for information interaction, a device, and a storage medium are provided. The method for information interaction includes: selecting a first target effect based on an effect selection operation input by a first user for a target object, where the target object is associated with a preset theme; playing the first target effect in response to receiving a preset playing instruction for the target object; and in response to meeting a preset condition, showing interaction task information associated with the preset theme.

Abstract: A compound represented by formula (I) and a preparation method thereof, and a pharmaceutically acceptable salt thereof, a metabolic precursor thereof, a metabolite thereof, an isomer thereof or a prodrug thereof; experiments show that the indole alkaloid not only promotes axon growth of peripheral sensory neurons, improves the nerve conduction velocity and anesthesia symptom of diabetic rats, but also promotes healing of foot ulcer wounds of diabetic rats, thus having good therapeutic effects on diabetic complication peripheral neuropathy and diabetic feet; they als show that the indole alkaloid also obviously reduces a degree of pulmonary fibrosis induced by bleomycin in mice, and plays a role in protecting lung tissues.

Abstract: One or more systems, devices, computer program products and/or computer-implemented methods of use provided herein relate to generating images from sound waves using discrete natural language tokens as bridging features. The computer-implemented system can comprise a memory that can store computer executable components. The computer-implemented system can further comprise a processor that can execute the computer executable components stored in the memory, wherein the computer executable components can comprise an image generation component that can use a first neural network model to generate an image of an environment detected by an imaging sonar, based on discrete tokens in natural language that can represent sound waves reflected by structures in the environment, wherein the discrete tokens can be non-semantic.

Abstract: Disclosed are various approaches for using drag-and-drop to open local files remotely. In one approach, a client device renders a user interface including a first indicium corresponding to a local file of the client device and a second indicium corresponding to a remotely executed application. A drag-and-drop action of the first indicium relative to the second indicium is detected. Responsive to the drag-and-drop action, the remotely executed application opens a copy of the local file.

Abstract: A service includes a trained model comprising a classifier that predicts whether domain names are dictionary DGA generated. Using passive DNS data and/or a heuristic analysis based on natural language processing of the domain name, the service filters domain names that are not candidate (i.e., potential) dictionary DGA domain names out of the detection pipeline. There domain names are thus classified without being fed into the model for more computationally expensive processing. Domain names that are not filtered out are queued for input into an instance of the model and classification by the model, with the queued domain names processed in small batches and load balanced across model instances. Predicted domain name classes output by the model are cached for subsequent cache reads to avoid multiple runs of the model for one domain name.

Abstract: A method of forming a thin film structure involves performing one or more repetitions to form a template on a wafer. The repetitions include: depositing a layer of a template material to a thickness; and ion beam milling the layer of the template material to remove thickness less than the first thickness. The ion beam milling may be performed at a two different angles during two different repetitions. At least one of the angles is a channeling angle defined relative to a crystalline microstructure of the template material. After the repetitions, additional material may be deposited on the template to form a final structure. The additional material has a same crystalline microstructure as the template material.

Abstract: A method, computer program product, and computer system are provided for predicting treatment options based on cardiac auscultation data. Text data and audio data corresponding to cardiac auscultation associated with a patient is received. The text data and the audio data are encoded as respective text vectors and audio vectors. A distance between the text vectors and the audio vectors is calculated. Diagnosis results are determined by a machine learning model based on the calculated distance between the text vectors and the audio vectors.

Abstract: According to one embodiment, a method, computer system, and computer program product for building an acoustic model is provided. The present invention may include performing contrastive pre-training of the acoustic model; building a dataset classifier using prompt engineering; performing a prediction process; and performing zero-shot audio prediction using the pre-trained acoustic model.

Abstract: Embodiments of the present disclosure provide systems and methods for implementing enhanced Optical Character Recognition (OCR) of text overlapping scenes through text graph structuring. Text graph structuring is performed to provide a graph data structure for each data character or letter of multiple letters and a library of graph templates from graph structured data of each of the multiple letters. Text graph structuring is performed to convert visual content of an identified overlapping text image region to an overlapping text topology graph. The overlapping text topology graph is split into multiple subgraphs using the graph template library to match recognizable letters in the overlapping text.