Abstract: A deep-sea crawling robot and a crawling method thereof are provided. The deep-sea crawling robot includes a crawling robot body and a countertop, wherein a top of the crawling robot body is fixedly connected to the countertop, left and right sides of a bottom of the crawling robot body are respectively provided with crawler belts, and left and right sides of the countertop are respectively provided with weight-increasing auxiliary crawling mechanisms; and an outer plate is fixedly connected to a front side or back side of the countertop, a multifunctional propelling mechanism is arranged on the outer plate, and a rapid dewatering mechanism is arranged in a chamber of the countertop. The deep-sea crawling robot can not only accelerate the propulsion and crawling speed when it is on the seabed but also further enhance its stability during the crawling process on the seabed.

Abstract: A bottom touching assisting device suitable for deep-sea submersibles and an implementation method thereof are provided. Four support columns are arranged within a mounting box body at a bottom of a deep-sea submersible, the support columns and the mounting box body are connected in a sliding manner through sliders, and the support column is sleeved with a threaded sleeve in a threaded connection manner. In conjunction with a drive component and a pressing mechanism, smooth vertical movement of the support column is achieved when the threaded sleeve rotates. This allows a bottom end of the support column to extend from the mounting box body.

Abstract: A wideband terahertz modulator based on gradual openings, which belongs to the technical field of electromagnetic functional devices, includes: a semiconductor substrate; an epitaxial layer provided on the semiconductor substrate; a modulation units array, a positive voltage loading electrode and a negative voltage loading electrode which are provided on the epitaxial layer; wherein each modulation unit in the modulation units array comprises a disconnected H-shaped structure, a metal electrode located below an end of the opening of the disconnected H-shaped structure, and a semiconductor doped heterostructure located below the opening of the disconnected H-shaped structure; wherein in the disconnected H-shaped structures, adjacent modulation units have different opening positions; in a same row, the opening positions are linearly distributed and have a certain slope, and inclination slopes of the opening positions of two adjacent rows are opposite.

Abstract: A computer device acquires speech content. The device performs feature extraction on the speech content to obtain an intermediate feature. The intermediate feature is used for indicating an audio expression characteristic of the speech content. The device decodes the intermediate feature based on an attention mechanism to obtain a first word graph network. The device performs feature mapping on the intermediate feature based on pronunciation of the speech content to obtain a second word graph network. The device determines a recognition result of the speech content according to candidate word connection relationships indicated by the first word graph network and the second word graph network.

Abstract: A speech recognition method includes acquiring speech data, inputting a speech feature matrix of the speech data to a speech recognition model, the speech feature matrix being used for representing time-domain and frequency-domain features of the speech data, performing attention encoding on the speech feature matrix through the speech recognition model to obtain an encoded matrix, the encoded matrix including multiple encoded vectors, and decoding the multiple encoded vectors in the encoded matrix according to positions of the multiple encoded vectors in the encoded matrix to output a character string corresponding to the speech data, a decoding sequence of the multiple encoded vectors being related to the positions of the multiple encoded vectors in the encoded matrix.

Abstract: A speech recognition method includes: obtaining speech information; and determining beginning and ending positions of a candidate speech segment in the speech information by using a weighted finite state transducer (WFST) network. The candidate speech segment is identified as corresponding to a preset keyword. The method also includes clipping the candidate speech segment from the speech information according to the beginning and ending positions of the candidate speech segment; detecting whether the candidate speech segment includes a preset keyword by using a machine learning model; and determining, upon determining that the candidate speech segment comprises the preset keyword, that the speech information comprises the preset keyword.

Abstract: A high-electron mobility transistor (HEMT) array terahertz wave modulator loaded in a waveguide is provided, which belongs to the technical field of electromagnetic functional devices and focuses on fast dynamic functional devices in the terahertz band. The device comprises a waveguide cavity and a modulation chip. The modulation chip comprises a semiconductor material substrate, a heterostructure material epitaxial layer, an artificial microstructure, and a socket circuit. The applied voltage controls the distribution change of the two-dimensional electron gas in the HEMT, which in turn controls the resonance mode conversion in the artificial microstructure, thereby control the transmission of electromagnetic waves in the waveguide. The modulator has a modulation depth of up to 96% and a modulation rate above 2 GHz.

Abstract: A speech recognition method includes: obtaining speech data; performing feature extraction on speech data, to obtain speech features of at least two speech segments; inputting the speech features of the at least two speech segments into the speech recognition model, and processing the speech features of the speech segments by using cascaded hidden layers in the speech recognition model, to obtain hidden layer features of the speech segments, a hidden layer feature of an ith speech segment being determined based on speech features of n speech segments located after the ith speech segment in a time sequence and a speech feature of the ith speech segment; and obtaining text information corresponding to the speech data based on the hidden layer features of the speech segments.

Abstract: A speech recognition method includes: obtaining speech information; and determining beginning and ending positions of a candidate speech segment in the speech information by using a weighted finite state transducer (WFST) network. The candidate speech segment is identified as corresponding to a preset keyword. The method also includes clipping the candidate speech segment from the speech information according to the beginning and ending positions of the candidate speech segment; detecting whether the candidate speech segment includes a preset keyword by using a machine learning model; and determining, upon determining that the candidate speech segment comprises the preset keyword, that the speech information comprises the preset keyword.

Abstract: A high-electron mobility transistor (HEMT) array terahertz wave modulator loaded in a waveguide is provided, which belongs to the technical field of electromagnetic functional devices and focuses on fast dynamic functional devices in the terahertz band. The device comprises a waveguide cavity and a modulation chip. The modulation chip comprises a semiconductor material substrate, a heterostructure material epitaxial layer, an artificial microstructure, and a socket circuit. The applied voltage controls the distribution change of the two-dimensional electron gas in the HEMT, which in turn controls the resonance mode conversion in the artificial microstructure, thereby control the transmission of electromagnetic waves in the waveguide. The modulator has a modulation depth of up to 96% and a modulation rate above 2 GHz. The invention can be realized by using micro-processing technology, and the preparation process is mature and reliable.

Abstract: A wideband terahertz modulator based on gradual openings, which belongs to the technical field of electromagnetic functional devices, includes: a semiconductor substrate; an epitaxial layer provided on the semiconductor substrate; a modulation units array, a positive voltage loading electrode and a negative voltage loading electrode which are provided on the epitaxial layer; wherein each modulation unit in the modulation units array comprises a disconnected H-shaped structure, a metal electrode located below an end of the opening of the disconnected H-shaped structure, and a semiconductor doped heterostructure located below the opening of the disconnected H-shaped structure; wherein in the disconnected H-shaped structures, adjacent modulation units have different opening positions; in a same row, the opening positions are linearly distributed and have a certain slope, and inclination slopes of the opening positions of two adjacent rows are opposite.