Abstract: Systems and methods are provided for obtaining, training, and using an end-to-end AST model based on a neural transducer, the end-to-end AST model comprising at least (i) an acoustic encoder which is configured to receive and encode audio data, (ii) a prediction network which is integrated in a parallel model architecture with the acoustic encoder in the end-to-end AST model, and (iii) a joint layer which is integrated in series with the acoustic encoder and prediction network. The end-to-end AST model is configured to generate a transcription in the second language of input audio data in the first language such that the acoustic encoder learns a plurality of temporal processing paths.

Abstract: A method for vehicle charge control includes: acquiring a proportional relationship between an estimated charging cost and an estimated charging time and a target charging quantity of electricity; determining the estimated charging cost and the estimated charging time according to the target charging quantity of electricity and the proportional relationship between the estimated charging cost and the estimated charging time; and determining an actual control sequence for performing an actual charge control on the vehicle according to the target charging quantity of electricity, the estimated charging cost and the estimated charging time, in which the actual control sequence is a sequence consisting of control signals issued to a component related to charging of the vehicle when the vehicle is charged.

Abstract: A computer implemented method includes receiving speech data representative of speech in a first language The speech data is divided into chunks of speech data, each chunk comprising multiple temporally consecutive frames of acoustic information. Each temporally consecutive chunk of data is processed using beam search on each frame to identify candidate language tokens representing a second language different from the first language. A best candidate language token(s) is selected for each chunk as processed. The selected best candidate language token or tokens for each chunk of data is committed as a prefix for a next temporally consecutive chunk of data.

Abstract: Systems and methods are provided for accessing a factorized neural transducer comprising a first set of layers for predicting blank tokens and a second set of layers for predicting vocabulary tokens, the second set of layers comprising a language model that includes a vocabulary predictor which is a separate predictor from the blank predictor, wherein a vocabulary predictor output from the vocabulary predictor and the encoder output are used for predicting a vocabulary token. The second set of layers is selectively modified to facilitate an improvement in an accuracy of the factorized neural transducer in performing automatic speech recognition, the selectively modifying comprising applying a particular modification to the second set of layers while refraining from applying the particular modification to the first set of layers.

Abstract: Provided is a method for preparing a crosslinked polyvinyl chloride structure foamed material by using a physical foaming agent, comprising: melting and mixing a polyvinyl chloride resin, a modified resin, an isocyanate, an anhydride, a nucleating agent and a heat stabilizer to obtain a blank; immersing the blank into a foaming gas for foaming to obtain a foamed body; and subjecting the foamed body to crosslinking curing to obtain a crosslinked polyvinyl chloride structure foamed material. According to the present invention, the crosslinked polyvinyl chloride structure foamed material is prepared by means of a brand-new process route, and the composition and proportion of the PVC resin and other raw materials can be adjusted in a wide range; and carbon dioxide or nitrogen is used for foaming. Further provided is a crosslinked polyvinyl chloride structure foamed material.

Abstract: This disclosure provides triptolide-conjugates, methods of making such compounds, pharmaceutical compositions and medicaments comprising such compounds, as well methods of using such compounds in the treatment of conditions/diseases, such as those relating to cancer, immunomodulation and/or inflammation.

Abstract: Systems and methods are provided for obtaining, training, and using an end-to-end AST model based on a neural transducer, the end-to-end AST model comprising at least (i) an acoustic encoder which is configured to receive and encode audio data, (ii) a prediction network which is integrated in a parallel model architecture with the acoustic encoder in the end-to-end AST model, and (iii) a joint layer which is integrated in series with the acoustic encoder and prediction network. The end-to-end AST model is configured to generate a transcription in the second language of input audio data in the first language such that the acoustic encoder learns a plurality of temporal processing paths.

Abstract: Disclosed herein is a herbicidal composition, comprising an active component A and an active component B, wherein the active component B is one selected from the group consisting of glufosinate-ammonium, glufosinate-P-ammonium, oxyfluorfen, acetoflufen, imazethapyr, dimethoxamide, dimethoxamid-P, metolachlor, S-metolachlor, butachlor, acetochlor, pyroxasulfone, glyphosate and derivatives thereof.

Abstract: Solutions for on-device streaming inverse text normalization (ITN) include: receiving a stream of tokens, each token representing an element of human speech; tagging, by a tagger that can work in a streaming manner (e.g., a neural network), the stream of tokens with one or more tags of a plurality of tags to produce a tagged stream of tokens, each tag of the plurality of tags representing a different normalization category of a plurality of normalization categories; based on at least a first tag representing a first normalization category, converting, by a first language converter of a plurality of category-specific natural language converters (e.g., weighted finite state transducers, WFSTs), at least one token of the tagged stream of tokens, from a first lexical language form, to a first natural language form; and based on at least the first natural language form, outputting a natural language representation of the stream of tokens.

Abstract: Provided are a functional coating material for a lithium ion battery separator and a preparation method therefor. The functional coating material for a lithium ion battery separator comprises 1-15% mass fraction of inorganic nanoparticles, 8-30% mass fraction of nanofibers and 1-5% mass fraction of an adhesive, and the remainder being a solvent. The preparation method therefor comprises the following steps: (1) preparing the nanofibers as a spinning precursor liquid; (2) preparing a dispersion liquid with the inorganic nanoparticles and the spinning precursor liquid; and (3) adding the adhesive to the dispersion liquid, and mixing same until uniform to obtain the functional coating material for a lithium ion battery separator.

Abstract: A polypeptide magnetic nanoparticle, comprising: a specific targeting polypeptide and a magnetic nanoparticle. The amino acid sequence of the specific targeting polypeptide is VRRDAPRFSMQGLDA-X, and the C-terminal X thereof is a sequence of 5 to 20 amino acids. The amino acid in the amino acid sequence of C-terminal X is selected from one or more of: C, G, and N. The polypeptide magnetic nanoparticle can be used for CTC detection and molecular typing of various cancer types, including esophageal cancer, liver cancer, lung cancer, stomach cancer, bladder cancer, skin cancer, melanoma, breast cancer, colorectal cancer, cervical cancer, etc. Biomarkers of CTC molecular typing include PD-L1, HER2, ER, PR, AR, EGFR, VEGFR, CXCR4, etc. The polypeptide magnetic nanoparticle is helpful for CTC detection and the qualitative and semiquantitative analysis of expression levels of biomarkers at CTC level, thereby performing precise therapy such as targeted individualized targeted therapy or immunotherapy on patients.

Abstract: A composite diaphragm, comprising first particles and second particles, and a binder less than 5 wt %. The first particles and the second particles have different particle sizes of particle compositions, the first particles have a radius of r, and the second particles have a radius of r?, wherein r and r? meet the following relationship: (2?{square root over (3)}/3?1)<(r?/r)<(?{square root over (6)}/2?1); the first particles are prepared from at least one or more of egg shells, duck egg shells, goose egg shells, and other bird or amphibian egg shells; the second particles are prepared from shells and abalone shells; a first particle layer, a second particle layer, and the binder are adopted to form a film, an integrated lithium ion battery composite diaphragm material is formed, an inorganic material is not easy to fall off, and a particle material plays a full role.

Abstract: A lithium ion battery diaphragm, comprising a substrate and a modification layer, wherein the modification layer is an inorganic coating; the inorganic coating is composed of a first particle layer and a second particle layer; the first particle layer has a first particle size r, and the second particle layer has a second particle size r?, wherein the particle sizes r and r? satisfy the following relationship: (2?{square root over (3)}/3?1)<(r?/r)<(?{square root over (6)}/2?1); first particles are at least one or more of boehmite, aluminum oxide, titanium oxide, calcium oxide, zinc oxide, copper oxide, and manganese oxide; and second particles are natural organic particles, and are prepared from natural organic shells, wherein the natural organic shells are selected from eggshells and seashells. The first particle layer and the second particle layer are combined to form an integrated lithium battery diaphragm material.

Abstract: Disclosed herein is a herbicidal composition, comprising an active component A and an active component B, wherein the active component B is one selected from the group consisting of glufosinate-ammonium, glufosinate-P-ammonium, oxyfluorfen, acetoflufen, imazethapyr, dimethoxamide, dimethoxamid-P, metolachlor, S-metolachlor, butachlor, acetochlor, pyroxasulfone, glyphosate and derivatives thereof.

Abstract: A heat dissipation structure (2) for a lighting device (1) is provided. The heat dissipation structure comprises at least two separate heat sinks (20, 10) for a light source (3) and a driver (7) for the light source, respectively. Each heat sink comprises fins (11, 21, 22) and a wall arrangement (15, 25). The at least two separate heat sinks are disposed along an axial direction of the lighting device. The fins of the at least two separate heat sinks are enclosed by the wall arrangements to form a chimney structure (30) arranged along (such as substantially parallel with) the axial direction of the lighting device. Further, the chimney structure comprises at least two sub-chimney structures (31, 32) arranged fluidly in parallel. The sub-chimney structures provide (accelerate) two air flows in parallel within the chimney structure, whereby cool air flows through the two sub-chimney structures simultaneously, which reduces heat concatenation.

Abstract: A heat dissipation structure (2) for a lighting device (1) is provided. The heat dissipation structure comprises at least two separate heat sinks (20, 10) for a light source (3) and a driver (7) for the light source, respectively. Each heat sink comprises fins (11, 21, 22) and a wall arrangement (15, 25). The at least two separate heat sinks are disposed along an axial direction of the lighting device. The fins of the at least two separate heat sinks are enclosed by the wall arrangements to form a chimney structure (30) arranged along (such as substantially parallel with) the axial direction of the lighting device. Further, the chimney structure comprises at least two sub-chimney structures (31, 32) arranged fluidly in parallel. The sub-chimney structures provide (accelerate) two air flows in parallel within the chimney structure, whereby cool air flows through the two sub-chimney structures simultaneously, which reduces heat concatenation.

Abstract: A heat dissipation structure (2) for a lighting device (1), comprising: at least two separate heat sinks (20, 10) for a light source (3) and a driver (7) for the light source (1), respectively, each heat sink (20, 10) comprising fins (21, 22, 11) and a wall arrangement (15, 25), wherein said at least two separate heat sinks (20, 10) are disposed along an axial direction of the lighting device (1), wherein the fins (21, 22, 11) of said at least two separate heat sinks (20, 10) are enclosed by the wall arrangements (15, 25) to form a chimney structure (30) arranged along said axial direction of the lighting device (1), and wherein the chimney structure (30) comprises at least two sub-chimney structures (31, 32) arranged fluidly in parallel to provide two air flows simultaneously, which reduces heat concatenation.

Abstract: A Deep Neural Network (DNN) model used in an Automatic Speech Recognition (ASR) system is restructured. A restructured DNN model may include fewer parameters compared to the original DNN model. The restructured DNN model may include a monophone state output layer in addition to the senone output layer of the original DNN model. Singular value decomposition (SVD) can be applied to one or more weight matrices of the DNN model to reduce the size of the DNN Model. The output layer of the DNN model may be restructured to include monophone states in addition to the senones (tied triphone states) which are included in the original DNN model. When the monophone states are included in the restructured DNN model, the posteriors of monophone states are used to select a small part of senones to be evaluated.

Abstract: A Deep Neural Network (DNN) model used in an Automatic Speech Recognition (ASR) system is restructured. A restructured DNN model may include fewer parameters compared to the original DNN model. The restructured DNN model may include a monophone state output layer in addition to the senone output layer of the original DNN model. Singular value decomposition (SVD) can be applied to one or more weight matrices of the DNN model to reduce the size of the DNN Model. The output layer of the DNN model may be restructured to include monophone states in addition to the senones (tied triphone states) which are included in the original DNN model. When the monophone states are included in the restructured DNN model, the posteriors of monophone states are used to select a small part of senones to be evaluated.

Abstract: A method for selecting a product to display on a webpage of an e-commerce website. Online user behavior data associated with a plurality of merchants selling products on the webpage is obtained. The behavior data is user actions performed online with respect to merchants. The behavior data is analyzed to determine a most relevant group of the users for each merchant. The most relevant group is compared to a targeted group of the webpage to determine a rate of overlap for each merchant. The rate of overlap represents a percentage of the most relevant group that matches the targeted group. A top merchant is selected based on the highest rate of overlap. A top product being sold by the top merchant is selected to display on the webpage. The top product is displayed on the webpage of the e-commerce web site to an end user.