Abstract: Systems and methods for efficient information retrieval for clinical data are provided. A document including one or more patient health records can be obtained and tokenized to generate a sequence of tokens, each corresponding to a word or sub-word within the document. A plurality of segments each containing a token sequence comprising a sub-set of the tokens are then generated. Next, a transformer model is used to process each of the plurality of segments to generate a corresponding word-level encoding for each segment. The word-level encodings for each of the segments are combined and fused to obtain document-level contextual data. The combined and fused word-level encodings can then be analyzed, such as to identify named entities and relationships between them.

Abstract: An example operation may include one or more of receiving a natural language input, executing a generative machine learning (ML) model on the natural language input to generate a structured query language (SQL) statement, executing the SQL statement on fake mockup data to generate a first query result on the fake mockup data, executing the generative ML model on the natural language input and the fake mockup data to generate a second query result on the fake mockup data, and determining whether the SQL statement is valid based on a comparison of the first query result and the second query result.

Abstract: The present invention provides a generation method and apparatus for an optimal transformation strategy for building energy, a device, and a medium. An urban building energy simulation result is obtained on the basis of multidimensional data, and a training sample is constructed to perform integrated training on a plurality of machine learning models to obtain a surrogate model, in this way, during subsequent building performance simulation, the surrogate model can be directly used for prediction, so that the processing efficiency is increased; and multi-objective optimization processing is performed on the basis of the surrogate model and a parallel genetic algorithm plugin, so that an optimal transformation strategy for a target building is rapidly generated. Not only is the quantification for a transformation result achieved, but also the generated transformation strategy is more reasonable.

Abstract: The present disclosure discloses a flexible mechanical arm and a robot. The flexible mechanical arm includes an arm joint, a flexible rotary module (2) and an end execution mechanism (3) which are connected with each other. The flexible rotary module (2) is driven by fluid, and includes a first end plate (23), a powered pusher plate (222), a first folding unit (21), a nut (221), and a lead screw (223) in threaded connection with the nut (221), wherein the first folding unit (21) is driven by fluid to be extended or compressed to push the powered pusher plate (222) to translate in an axial direction of the lead screw (223). The flexible mechanical arm can be deformed under drive of the fluid, thus causing no damage to surrounding objects or persons, and capable of being applied to special environments such as electric fields, magnetic fields, rays, etc.

Abstract: A circuit includes a receiver coupled to an input node and includes a transimpedance amplifier (TIA). The input node is to be coupled to a transmitter over a channel. Calibration circuitry is coupled to the input node and to modify a resistance of a feedback resistor of the receiver by applying a plurality of control values. The calibration circuitry is to generate, for each applied control value, a first digital code for a direct current (DC) swing and a second digital code for an alternating current (AC) swing at the input node. The calibration circuitry is to determine a final control value of the plurality of control values that minimizes a difference between the first and second digital codes.

Abstract: A method and a system for fully automatically segmenting a cerebral cortex surface based on a graph network, the method including: registering a cerebral magnetic resonance image to a standard template space; performing cerebral cortex surface reconstruction on the registered cerebral magnetic resonance image based on a deep neural network; calculating adjacency matrices among the vertexes of the reconstructed cerebral cortex; acquiring corresponding distinguishing features of each grid vertex of the cerebral cortex surface as corresponding feature vectors of the point so as to obtain a cerebral cortex surface segmentation result; and mapping the cerebral cortex surface segmentation result back to the original coordinate space from the standard template space. Graph structure modeling is performed on the reconstructed cerebral cortex surface, and the global topological structure features are learned based on the graph network, thus realizing the accurate segmentation of the cerebral cortex surface.

Abstract: The present disclosure provides a corrugated plate and a liquefied gas storage container having the same. The corrugated plate comprises a plate body, a first corrugation and a second corrugation orthogonal to each other, and an intersection structure formed at the intersection position. The first corrugations are higher than the second corrugations. The intersection structure comprises a top portion, ridges, and recessed portions. The recessed portions consist of four, each located in one quadrant, formed on the outer side of the first corrugation and sandwiched between a first crest of the first corrugation and a corresponding ridge, and concaved inward relative to the outer side. According to the present disclosure, the intersection structure ensures that the upper and lower surfaces at the intersection position have uniform line lengths, thereby achieving uniform thickness, avoiding excessive thinning, and providing improved structural strength and fatigue resistance.

Abstract: A circuit includes a receiver coupled to an input node and includes a transimpedance amplifier (TIA). The input node is to be coupled to a transmitter over a channel. Calibration circuitry is coupled to the input node and to modify a resistance of a feedback resistor of the receiver by applying a plurality of control values. The calibration circuitry is to generate, for each applied control value, a first digital code for a direct current (DC) swing and a second digital code for an alternating current (AC) swing at the input node. The calibration circuitry is to determine a final control value of the plurality of control values that minimizes a difference between the first and second digital codes.

Abstract: The present disclosure provides a pea peptide with auxiliary hypoglycemic function, and preparation method and application thereof. The pea peptide includes at least peptide segments pEE, pEK and pER in its composition; and based on a mass of the pea peptide, a content of the peptide segment pEE is ?100.00 mg/100 g, a content of the peptide segment pEK is ?80.00 mg/100 g and a content of the peptide segment pER is ?90.00 mg/100 g. The pea peptide has a significant efficacy in an aspect of reducing blood glucose.

Abstract: The present disclosure provides a corrugated plate and a liquefied gas storage container having the same. The corrugated plate comprises a plate body, a first corrugation and a second corrugation orthogonal to each other, and an intersection structure formed at the intersection position. The first corrugations are higher than the second corrugations. The intersection structure comprises a top portion, ridges, and recessed portions. The recessed portions consist of four, each located in one quadrant, formed on the outer side of the first corrugation and sandwiched between a first crest of the first corrugation and a corresponding ridge, and concaved inward relative to the outer side. According to the present disclosure, the intersection structure ensures that the upper and lower surfaces at the intersection position have uniform line lengths, thereby achieving uniform thickness, avoiding excessive thinning, and providing improved structural strength and fatigue resistance.

Abstract: A circuit includes a phase selector to generate an injection clock signal having an injection phase based on a phase of a digitally controlled oscillator clock signal generated within a phase-locking feedback loop. An injection-locked oscillator (ILO), coupled to an output of the phase selector, generates an ILO clock signal that is convertible to provide a feedback clock signal of the circuit. Logic, coupled between an output of the ILO and the phase selector, to, at each predetermined number of cycles of the DCO clock signal, cause the phase selector to output a phase shift in the injection clock signal that causes the ILO clock signal to comprise a rotated phase, relative to the injection phase, and that prevents a glitch in the injection clock signal.

Abstract: A circuit includes a phase selector to generate an injection clock signal having an injection phase based on a phase of a digitally controlled oscillator clock signal generated within a phase-locking feedback loop. An injection-locked oscillator (ILO), coupled to an output of the phase selector, generates an ILO clock signal that is convertible to provide a feedback clock signal of the circuit. Logic, coupled between an output of the ILO and the phase selector, to, at each predetermined number of cycles of the DCO clock signal, cause the phase selector to output a phase shift in the injection clock signal that causes the ILO clock signal to comprise a rotated phase, relative to the injection phase, and that prevents a glitch in the injection clock signal.

Abstract: An oyster peptide with an effect of improving sexual function and a preparation method thereof are provided, the oyster peptide at least includes peptide segments RI, IR and VR in its composition. Based on a mass of the oyster peptide, a content of the RI is ?3.60 mg/100 g, a content of the IR is ?7.60 mg/100 g, and a content of the VR is ?6.50 mg/100 g.

Abstract: A communications circuit includes a receiver (RX) frontend coupled to an RX deserializer and the RX frontend includes a decision feedback equalizer coupled directly between an output and an input of the RX frontend. The decision feedback equalizer may include a first inverter and a series combination of a first resistor and a first capacitor coupled in series with the first inverter.

Abstract: Embodiments of this specification provide language model training methods and apparatuses based on continual pre-training. In one example method, a soft prompt feature corresponding to each current training sample in a current domain is obtained. A latent feature corresponding to each piece of text data in the current domain is obtained. A cross-domain loss value is determined based on a difference between the obtained latent feature and a latent feature obtained based on an initial current language model in a previous domain. When a training termination condition for the current domain is not satisfied, model parameters are adjusted based on the cross-domain loss value. When a training termination condition for the current domain is satisfied, the model training process is repeated by continuing using a training sample set in a next domain, until a training termination condition for continual pre-training is satisfied.

Abstract: Provided is an automatic event graph construction method for multi-source vulnerability information. The method includes the following steps. A vulnerability report is crawled from a vulnerability database, a cause of vulnerability is taken as an event trigger word, and a vulnerability type is determined through the cause of vulnerability. An attacker, consequence, location and other information in a description are identified by named-entity recognition, and information completion is performed. An explicit relation between events is extracted by using text information, an implicit relation between events is extracted by using text similarity, and vulnerability-related code representation is performed. Obtained vulnerability event information is visualized into an event graph through a visualization tool.

Abstract: Disclosed are an integrated installation structure for a marine apparatus, a storage container and a manufacturing method. The integrated installation structure includes a primary insulating layer formed by arranging at least one first primary insulating plate and at least one second primary insulating plate, a secondary insulating layer disposed on an outer side of the primary insulating layer and a secondary shielding film between the primary insulating layer and the secondary insulating layer. The first primary insulating plate is prefabricated and connected with the secondary shielding film into a one-piece member through a first connecting device, and the second primary insulating plate is a structure separated from the secondary shielding film, and is connected with the secondary insulating layer through a second connecting device.

Abstract: The present invention provides a welding apparatus and a welding method for ultra-narrow welding gap, relating to the technical field of welding with an ultra-narrow welding gap. The welding apparatus for ultra-narrow welding gap comprises: a blowing mechanism, a conductive nozzle and an industrial water-cooling unit. The industrial water-cooling unit is connected to the blowing mechanism. A medium flow of the industrial water-cooling unit flows through the inside of the blowing mechanism. The industrial water-cooling unit is configured to reduce a temperature of the shielding gas outputted by the blowing mechanism and a temperature of the conductive nozzle.

Abstract: Cross-linked shear-thinning fluids of pectic acid demonstrating increased viscosity with decreasing shear, as well as methods of producing and using the same. A shear-thinning fluid includes an aqueous solution of pectic acid cross-linked by a divalent cation is disclosed. The pectic acid may be present in an amount ranging from about 0.5 to about 3.0% (w/v), the divalent cation may be present at a concentration of from about 0.5 mM to about 7.0 mM, and the viscosity of the shear-thinning fluid increases with decreasing shear. These cross-linked shear-thinning fluids of pectic acid can be utilized for controlled release formulations, cell encapsulation, and 3D printing, and provide for an improved balance between structure fidelity/mechanical stability and cell viability.

Abstract: A global heat flow topology optimization tool for cold energy recovery of floating storage regasification units is provided, including: comparing heat flow topology based on energy conservation of each component of a system and inlet and outlet conditions, and establishing a preliminary heat flow topology tool; based on Kirchhoff's law and thermal current method, eliminating a redundant heat flow loop and establishing the final global heat flow topology tool; determining heat flow balance, thermal-hydraulic balance, working medium types, system working conditions and various constraint conditions of coupling links of each system; solving the heat flow topology tool and comprehensively analyzing the system performances; determining the objective functions, setting the decision-making interval, performing multi-objective optimization based on intelligent algorithm, and solving the optimal working condition parameters of the system; and introducing a system example, and analyzing the above methods to verify the