Abstract: A term weight determination method includes: obtaining a video and video-associated text, the video-associated text including at least one term; generating a halfway vector of the term by performing multimodal feature fusion on the features of the video, the video-associated text and the at least one term; and generating the weight of the at least one term based on the halfway vector of the at least one term.

Abstract: A structure to apply flexing to an object to test resilience to being bent has a first portion and a second portion. The structure includes a base, a rotating member, a pressing plate, and an adjusting assembly. The base defines first and second receiving grooves, and the second receiving groove receives the first portion. The rotating member on the base is partially received in the first receiving groove. An abutting surface of the rotating member is recessed to form a third receiving groove which receives the second portion. The pressing plate on the base presses against the first portion. The adjusting assembly controls the rotating member to rotate towards the base about a desired angle, thereby causing the second portion to be bent with respect to the first portion.

Abstract: A voice identity feature extractor training method includes extracting a voice feature vector of training voice. The method may include determining a corresponding I-vector according to the voice feature vector of the training voice. The method may include adjusting a weight of a neural network model by using the I-vector as a first target output of the neural network model, to obtain a first neural network model. The method may include obtaining a voice feature vector of target detecting voice and determining an output result of the first neural network model for the voice feature vector of the target detecting voice. The method may include determining an I-vector latent variable. The method may include estimating a posterior mean of the I-vector latent variable, and adjusting a weight of the first neural network model using the posterior mean as a second target output, to obtain a voice identity feature extractor.

Abstract: A bone conduction microphone is provided. The bone conduction microphone may include a laminated structure formed by a vibration unit and an acoustic transducer unit. The bone conduction microphone may include a base structure configured to carry the laminated structure. At least one side of the laminated structure may be physically connected to the base structure. The base structure may vibrate based on an external vibration signal. The vibration unit may be deformed in response to the vibration of the base structure. The acoustic transducer unit may generate an electrical signal based on the deformation of the vibration unit. The bone conduction microphone may include at least one damping structural layer. The at least one damping structural layer may be arranged on an upper surface, a lower surface, and/or an interior of the laminated structure, and the at least one damping layer may be connected to the base structure.

Abstract: The present application provides an array substrate, a manufacturing method thereof, and a display panel. A display area of the array substrate includes a functional area, a sensing unit is arranged on a part of a base substrate corresponding to the functional area, the sensing unit is aligned with a pixel electrode, and a metal-covered area and a non-metal-covered area can form a same voltage difference with a common electrode, so that a deflection angle of each part of liquid crystal molecules in the functional area is consistent, and a problem of a difference between brightness and darkness of the display panel is solved.