Abstract: Techniques for processing a natural language query using an SMT solver that includes an LLM. The LLM processes the query text to formalize constraint text into pseudo code, which is processed by an SAT solver to determine logical atoms and propositional model for solving the query. The LLM then acts as a theory solver within the SMT solver to process the logical atoms and determine a valid solution for the natural language query.

Abstract: This disclosure provides an online test method and apparatus. The method includes obtaining a test question library, where the test question library includes a plurality of collected test questions. The method also includes obtaining a test model based on the test question library and a policy optimization algorithm, where the test model may be used to select at least one test question from the test question library in an online test process, the test model may include a state encoder and a recommender, the state encoder is configured to obtain a difference between input test questions to generate a state code, the recommender may be configured to output a test question based on the state code and an optimization objective, the optimization objective includes at least one of novelty or diversity.

Abstract: Methods, systems, and computer-readable storage media for a ML system that reduces a number of target items from consideration as potential matches to a query item using token embeddings and a search tree.

Abstract: A brassiere configured for gathering mamma accessoria includes a brassiere main body. The brassiere main body includes a first pressure portion, a second pressure portion, and a third pressure portion. The brassiere main body further includes two cups, two side wings, and two back wings. Outer side edges of the two cups are connected to the side wings. The two side wings are connected to the back wings. The first pressure portion is arranged at a junction between the side wing and the back wing. The second pressure portion is arranged on the side wing. The third pressure portion is arranged on the cup and is arranged close to the side wing.

Abstract: Recommendation system and method for cutter and coating combination based on workpiece parameters are provided.

Abstract: A method includes receiving an input audio signal that corresponds to utterances spoken by multiple speakers. The method also includes processing the input audio to generate a transcription of the utterances and a sequence of speaker turn tokens each indicating a location of a respective speaker turn. The method also includes segmenting the input audio signal into a plurality of speaker segments based on the sequence of speaker tokens. The method also includes extracting a speaker-discriminative embedding from each speaker segment and performing spectral clustering on the speaker-discriminative embeddings to cluster the plurality of speaker segments into k classes. The method also includes assigning a respective speaker label to each speaker segment clustered into the respective class that is different than the respective speaker label assigned to the speaker segments clustered into each other class of the k classes.

Abstract: A computer-implemented method of engineering design includes performing a simulation for a design under test of the design interacting in a real-world environment, in a computer processor, generating simulation data from performance of the simulation, storing the generated simulation data in a computer memory, extracting the generated simulation data to train a first inverse model neural network, and generating a plurality of design parameters from the inverse model neural network. A visualization representative of the generated plurality of design parameters for display to a user allows an expert user to evaluate the suggested design parameters and select some or all of the suggested design parameters. For existing designs more data is available for the simulation than for a new design. The inverse model for a new design may be augmented by transferring stored knowledge in a pre-existing inverse model to the model for the new design.

Abstract: A system and method for query selection are provided. The method may include acquiring a natural language sentence, pre-processing to obtain a standard node sequence, constructing a node tree based on the relationship between an index node and other nodes, generating a data query command based on the node tree, querying data using the data query command, and filtering the results. The standard node sequence may include at least an index node and a condition node. The node tree may be used to characterize the index-condition combination. The system may include an acquisition unit, a pre-processing unit, a node tree construction unit, a translation unit, and a querying and filtering unit.

Abstract: There are provided a touch display panel and a touch display device, the touch display panel includes: a display substrate including a plurality of pixel regions and a spacer region spacing the plurality of pixel regions apart from each other; a black matrix located on a light-exiting side of the display substrate, an orthographic projection of the black matrix on the display substrate being located in the spacer region; and a touch control structure layer including a first touch pattern layer located on the light-exiting side of the display substrate, the first touch pattern layer including a plurality of first metal lines, where an orthographic projection of each first metal line on the display substrate is located in the orthographic projection of the black matrix on the display substrate, and the first metal lines are in contact with the black matrix.

Abstract: The present disclosure provides a valve device having a housing and a valve core. The housing internally is provided with at least two fluid passages and sealing members located at a port of each fluid passage. The valve core is disposed in the housing and comprises at least one closing wall. The valve core is configured to be capable of rotating around an axis in a first direction in the housing to enable the closing wall to align with the corresponding fluid passage to close the aligned fluid passage. The valve core is configured to be capable of making a linear motion in a second direction in the housing to allow the closing wall to tightly abut against the sealing member at the closed fluid passage. The second direction forms an angle with an imaginary bisecting plane of the closing wall of the valve core, the angle is an acute angle.

Abstract: A method for preparing an environmental-friendly rubber oil, including: conducting extraction in an extraction tower, where a primary extraction solvent is introduced into an upper section of the extraction tower, a raw oil is introduced into a middle-lower section of the extraction tower, and a secondary extraction solvent is introduced into a lower section of the extraction tower; subjecting a material at a bottom of the extraction tower to solvent recovery to obtain a secondary extract oil; mixing a substance flowing out from a top of the extraction tower, a solubilizer and a part of the primary extraction solvent, cooling to ?15° C. to 10° C., and filtering to obtain a wax paste and a filtrate; subjecting the wax paste to heating evaporation and solvent recovery to obtain a wax; and subjecting the filtrate to heating evaporation and solvent recovery, obtaining the environmental-friendly rubber oil.

Abstract: A method for dewaxing of an extract oil and a rubber oil is provided, where a composite solvent including a polar solvent and a solubilizer is used as a dewaxing solvent, the polar solvent is N-methylpyrrolidone (NMP), and the solubilizer is at least one selected from the group consisting of benzene, toluene, and xylene. The method includes: mixing the composite solvent with the extract oil or the rubber oil at a temperature higher than a precipitation temperature of a wax, subjecting a resulting mixture to cooling crystallization and filtration to obtain a wax paste and a filtrate, and then independently subjecting the wax paste and the filtrate to solvent recovery by heating evaporation to obtain a wax and a dewaxed oil, respectively.

Abstract: An ophthalmic treatment device for photobiomodulation of a retina comprising: a treatment light source producing a continuous wave or quasi-continuous wave output beam having a wavelength in the range of 600 nm to 1000 nm and a power in the range of 1 mW to 500 mW; a beam homogenizing module that homogenizes the output beam of the treatment light source; a beam shaping module that modifies the output beam profile to produce a treatment light beam with an annular light beam profile; and a beam delivery and viewing module that allows an Operator to observe and operate to deliver the treatment light beam to the treatment location on the retina with an intensity in the range of 1 mW/cm2 to 500 mW/cm2.

Abstract: A flexible substrate, including: a first organic layer, an inorganic barrier layer, a metal layer and a second organic layer which are stacked in sequence, and the metal layer is for improving binding force between the inorganic barrier layer and the second organic layer. According to embodiments of the present disclosure, the metal layer is provided between the inorganic barrier layer and the second organic layer, so as to improve bonding force between the inorganic barrier layer and the second organic layer, prevent the second organic layer from peeling off from the inorganic barrier layer, and further improve reliability of a flexible display device.

Abstract: A method includes receiving an input audio signal that corresponds to utterances spoken by multiple speakers. The method also includes processing the input audio to generate a transcription of the utterances and a sequence of speaker turn tokens each indicating a location of a respective speaker turn. The method also includes segmenting the input audio signal into a plurality of speaker segments based on the sequence of speaker tokens. The method also includes extracting a speaker-discriminative embedding from each speaker segment and performing spectral clustering on the speaker-discriminative embeddings to cluster the plurality of speaker segments into k classes. The method also includes assigning a respective speaker label to each speaker segment clustered into the respective class that is different than the respective speaker label assigned to the speaker segments clustered into each other class of the k classes.

Abstract: Disclosed are a display substrate, a preparation method therefor, and a display device. The display substrate includes a display region and a binding region on one side of the display region. The binding region includes a binding structure layer disposed on a base. The binding structure layer includes a composite insulating layer disposed on the base. The binding region further includes a step structure formed by the base and the composite insulating layer. Heights of steps in the step structure decrease sequentially in the direction away from the display region. In the step structure, the base forms a first step having the smallest height. The binding structure layer further includes a signal connection wire having at least a portion thereof disposed on the step structure and located on the first step. An opening exposing the signal connection wire is provided on the base at the first step.

Abstract: A display substrate includes: a flexible light-transmitting substrate; a light-shielding layer disposed on a side of the flexible light-transmitting substrate; and a plurality of sub-pixels disposed on a side of the light-shielding layer away from the flexible light-transmitting substrate. The side of the flexible light-transmitting substrate has grooves. The light-shielding layer includes a plurality of light-shielding patterns, and a light-shielding pattern is located in a groove. The light-shielding pattern includes a first sub-light-shielding pattern covering a side wall of the groove and a second sub-light-shielding pattern covering a bottom surface of the groove. Each sub-pixel includes a pixel driving to circuit including a plurality of transistors. An active layer of a transistor is located in the groove.

Abstract: Methods, systems, and computer-readable storage media for receiving a first set of predictions generated by a ML model during execution of a training pipeline to train the ML model, each prediction in the first set of predictions being associated with a confidence, determining a set of confidence bins based on confidences of the first set of predictions, for each confidence bin in the set of confidence bins, providing an accuracy, processing the set of confidence bins and accuracies through a regression model to provide one or more regressions, each regression representing a confidence-to-accuracy relationship, defining a set of confidence thresholds based on at least one regression of the one or more regressions, and during an inference phase, applying the set of confidence thresholds to selectively filter predictions from a second set of predictions generated by the ML model.

Abstract: A data processing method includes obtaining feature representations of a plurality of knowledge points and a first learning state of a user, the first learning state indicates a user's level of mastery of a learned knowledge point, and the plurality of knowledge points are different from the learned knowledge point; and obtaining, by using a decoder based on the feature representations of the plurality of knowledge points, the first learning state, and a learning objective, a learning order corresponding to the plurality of knowledge points, the decoder is configured to identify a relationship between the feature representations of the plurality of knowledge points and the learning objective in the first learning state. The learning order of the plurality of knowledge points is determined in a sequence generation manner based on a current learning state of the user.