Abstract: A method for training a decision-making model parameter, a decision determination method, an electronic device, and a non-transitory computer-readable storage medium are provided. In the method, a perturbation parameter is generated according to a meta-parameter, and first observation information of a primary training environment is acquired based on the perturbation parameter. According to the first observation information, an evaluation parameter of the perturbation parameter is determined. According to the perturbation parameter and the evaluation parameter thereof, an updated meta-parameter is generated. The updated meta-parameter is determined as a target meta-parameter, when it is determined, according to the meta-parameter and the updated meta-parameter, that a condition for stopping primary training is met.

Abstract: The present disclosure proposes a language generation method and apparatus. The method includes: performing encoding processing on an input sequence by using a preset encoder to generate a hidden state vector corresponding to the input sequence; in response to a granularity category of a second target segment being a phrase, decoding a first target segment vector, the hidden state vector, and a position vector corresponding to the second target segment by using N decoders to generate N second target segments; determining a loss value based on differences between respective N second target segments and a second target annotated segment; and performing parameter updating on the preset encoder, a preset classifier, and the N decoders based on the loss value to generate an updated language generation model for performing language generation.

Abstract: There is provided a method and apparatus for generating a competition commentary based on artificial intelligence, and a storage medium. The method comprises: obtaining commentator's words commentaries and structured data of historical competitions; generating a commentating model according to obtained information; during live broadcast of a competition, determining a corresponding words commentary according to the commentating model with respect to the structured data obtained each time.

Abstract: The present disclosure provides a pre-training method for a sentiment analysis model and an electronic device, which relates to a field of artificial intelligence technologies. The method includes: based on a given seed sentiment dictionary, performing sentimental knowledge detection on a training corpus in a training corpus set, and determining a detection sentiment word and a detection word pair of the training corpus; according to preset mask processing rules, performing mask process on the training corpus to generate a masked corpus; performing encoding and decoding on the masked corpus by using a preset encoder and decoder to determine the detection sentiment word and the detection word pair of the training corpus; and updating the preset encoder and decoder according to a difference between prediction sentiment word and the detection sentiment word, and a difference between prediction word pair and the detection word pair.

Abstract: A delivery guide wire (10,100) and a therapeutic device are disclosed. The therapeutic device includes the delivery guide wire (10, 100), a medical implant and a delivery catheter (20). The delivery guide wire (10, 100) includes a core shaft (110) and a driving member (120) disposed on the core shaft (110), and the driving member (120) defines thereon a depression. The medical implant is compressed by the delivery catheter (20) and disposed over the delivery guide wire (10,100) in such a manner that it is at least partially received in the depression. This results in an increased contact area between the medical implant and the delivery guide wire (10, 100), which facilitates movement of the medical implant in sync with the delivery guide wire (10, 100) and makes its delivery easier.

Abstract: The present disclosure provides a method, apparatus, electronic device and storage medium for processing a semantic representation model, and relates to the field of artificial intelligence technologies. A specific implementation solution is: collecting a training corpus set including a plurality of training corpuses; training the semantic representation model using the training corpus set based on at least one of lexicon, grammar and semantics. In the present disclosure, by building the unsupervised or weakly-supervised training task at three different levels, namely, lexicon, grammar and semantics, the semantic representation model is enabled to learn knowledge at levels of lexicon, grammar and semantics from massive data, enhance the capability of universal semantic representation and improve the processing effect of the NLP task.

Abstract: Various embodiments of the teachings herein include a fault processing method comprising: receiving two historical faults similar to a target fault; searching keywords in a description of the target fault and each historical fault, wherein the keywords are classified into N grades, and for each system component in a grade, the grade comprises at least one keyword for describing the component, wherein N is an integer no less than 2; for each of the N grades, counting a quantity of identical system components represented by the keywords in the text description of each historical fault and the target fault; and comparing a degree of similarity of each historical fault to the target fault according to the quantity of identical system components counted in each grade of the N different grades, wherein a historical fault relating to a larger number of high-grade identical system components has a higher degree of similarity to the target fault.

Abstract: The present disclosure discloses a method and an apparatus for generating a text based on a semantic representation and relates to a field of natural language processing (NLP) technologies. The method for generating the text includes: obtaining an input text, the input text comprising a source text; obtaining a placeholder of an ith word to be predicted in a target text; obtaining a vector representation of the ith word to be predicted, in which the vector representation of the ith word to be predicted is obtained by calculating the placeholder of the ith word to be predicted, the source text and 1st to (i?1)th predicted words by employing a self-attention mechanism; and generating an ith predicted word based on the vector representation of the ith word to be predicted, to obtain a target text.

Abstract: A vascular implant, a delivery device and a medical apparatus are disclosed. The vascular implant includes a first engagement structure, and the delivery device includes the delivery shaft and a chamber. The delivery shaft includes a second engagement structure configured for detachable retaining engagement with the first engagement structure. Both the first engagement structure and the delivery shaft are configured to be received in the chamber. When the first engagement structure and the delivery shaft are received in the chamber, the first engagement structure is confined by the chamber and thus remains in retaining engagement with the second engagement structure, thus axially locking the vascular implant to the delivery shaft. When the first engagement structure is removed from the chamber, it is no longer confined thereby and is thus detachable from the second engagement structure, thus axially unlocking the vascular implant from the delivery shaft.

Abstract: A method for generating a target object is provided. A first discrete encoded sequence corresponding to an original object is generated by performing discrete encoding on the original object. The original object is of an image type, a text type, or a text-image-combined type. A second discrete encode sequence is obtained by inputting the first discrete encoded sequence into a generative model. A target object is generated based on the second discrete encoded sequence. The target object is of an image type or a text type. When the original object is of the image type, the target object is of the text type. When the original object is of the text type, the target object is of the image type.

Abstract: The present application provides a method of training a natural language processing model, which relates to a field of artificial intelligence, and in particular to a field of natural language processing. A specific implementation scheme includes: performing a semantic learning for multi-tasks on an input text, so as to obtain a semantic feature for the multi-tasks, wherein the multi-tasks include a plurality of branch tasks; performing a feature learning for each branch task based on the semantic feature, so as to obtain a first output result for each branch task; calculating a loss for each branch task according to the first output result for the branch task; and adjusting a parameter of the natural language processing model according to the loss for each branch task. The present application further provides a method of processing a natural language, an electronic device, and a storage medium.

Abstract: Various embodiments include a method for labeling a data point comprising executing a labeling operation on a target data set, wherein the target data set comprises a plurality of data points, each data point representing a service instance. The labeling operation comprises dividing the target data into subsets. For each subset, then: receiving input designating a mark for a first data point, illustrating the situation of the service instance represented by the data point; determining whether the similarity between the mark and a mark previously designated for a second data point in the target data set satisfies a preset condition; if the condition is not satisfied, taking the first subset as a target data set to re-execute the labeling operation; and if the condition is satisfied, setting, for each data point, a mark associated with the mark previously designated for a data point in the target data set.

Abstract: Embodiments of the present disclosure disclose a method and apparatus for subscribing to a topic. The method includes: matching a persistent topic for a retrieval keyword based on the retrieval keyword of a user combined with at least one of historical behavior or a subscription record of the user; returning the persistent topic to a client for display, so that the user performs subscription; and saving the persistent topic subscribed to by the user, and when a matching resource corresponding to the persistent topic subscribed to by the user is updated, recommending the updated matching resource to the user. Since a recommendation strategy can be timely adjusted according to real-time behavior of a user combined with historical behavior, the probability of hitting a topic of interest to the user is increased.

Abstract: A memory of a network device is divided into first blocks, each first block being divided into second blocks, and each second block including a first storage space and second storage space. When a packet is stored, second blocks occupied by the packet are determined based on a packet length and a first storage space length, and the packet is stored into a first storage space of each of the second blocks. For each of the second blocks, a PD corresponding to the second block is generated, and stored into a second storage space of the second block. When a packet is read, the second blocks to be read are determined based on a start address of the packet. A packet fragment is read from a first storage space of the second blocks to be read, and the read packet segments are composed into the second packet to be sent.

Abstract: A method for interactions in a live broadcast room includes: in response to opening a question interaction function in a live broadcast room, displaying a question panel interface in a first display area of the live broadcast room, wherein the question panel interface comprises a display card for each piece of question information and each display card comprises an answer control and an end control; and in response to triggering the answer control, displaying the target display card on top of remaining display cards on the question panel interface, and replacing the answer control with the end control.

Abstract: A power failure protection system is provided for a solid state drive. In an embodiment, the system includes: power converters; a stand-by power converter; a power switch array including power switch pairs; and a power control complex programmable logic device. The power control complex programmable logic device monitors power failures of the power converters; when a power failure of a particular power converter is detected, configures a power level for the stand-by power converter, the configured power level corresponding to the failed power converter; and controls a target power switch pair corresponding to the failed power converter such that an operation voltage generated by the stand-by power converter is applied to an internal component.

Abstract: A model training method, a model training platform, an electronic device and a storage medium are provided, which can be used in the field of artificial intelligence, particularly the fields of natural language processing and deep learning. The model training method includes: receiving an input; determining, based on the input, a user-oriented prefabricated function; determining, based on the input, a model training function; determining, based on the input, a pre-trained model; determining, based on the input, a network structure associated with the pre-trained model so as to support use of the pre-trained model; training, based on the input, the model by using the prefabricated function, the model training function, and the pre-trained model; and providing an output associated with a trained model.

Abstract: Presented herein are systems, methods, and datasets for automatically and precisely generating highlight or summary videos of content. In one or more embodiments, the inputs comprise a text (e.g., an article) of the key event(s) (e.g., a goal, a player action, etc.) in an activity (e.g., a game, a concert, etc.) and a video or videos of the activity. In one or more embodiments, the output is a short video of an event or events in the text, in which the video may include commentary about the highlighted events and/or other audio (e.g., music), which may also be automatically synthesized.

Abstract: Presented herein are systems, methods, and datasets for automatically and precisely generating highlight or summary videos of content. For example, in one or more embodiments, videos of sporting events may be digested or condensed into highlights, which will dramatically benefit sports media, broadcasters, video creators or commentators, or other short video creators, in terms of cost reduction, fast, and mass production, and saving tedious engineering hours. Embodiment of the framework may also be used or adapted for use to better promote sports teams, players, and/or games, and produce stories to glorify the spirit of sports or its players. While presented in the context of sports, it shall be noted that the methodology may be used for videos comprising other content and events.

Abstract: A method and an apparatus for interactions in a live broadcast room are disclosed. The method may include: in response to an opening operation for a question interaction function in a live broadcast room: obtaining at least one piece of question information from a server and displaying the at least one piece of question information in a first display area of the live broadcast room, in which the at least one piece of question information is sent by at least one audience client to the server; obtaining an interaction parameter value corresponding to each piece of question information from the server; and adjusting a position of the question information in the first display area based on the interaction parameter value corresponding to each piece of question information.