4 results about "Speech identification" patented technology

The invention discloses a voice recognition system and method based on audiovisual bimodal perception, and relates to the technical field of voice recognition, and the method comprises the steps: carrying out the feature recursion of an aligned bimodal sequence based on an audiovisual frame package, obtaining an audio high-level representation and a visual high-level representation, and calculating the modal reliability; and fusing the audio high-level representation and the visual high-level representation based on the modal reliability, and evaluating the prefix stability and confidence state through a streaming decoder to obtain a streaming identification information packet. According to the method, the audio high-level representation and the visual high-level representation are fused based on the modal reliability, and the streaming decoder is used for evaluating the prefix stability and the confidence state to obtain the streaming identification information packet, so that the decoding output has the confirmable stable fragment and the credibility representation, and the continuity and the consistency of the speech identification text are improved.

The present application discloses a synthesized speech identification method, apparatus and system, a storage medium, and a device. The method comprises: constructing a target data set on the basis of real speech data and synthesized speech data of multiple speakers, wherein each piece of speech data has a corresponding speaker label; constructing an identification model, wherein the identification model comprises a feature extraction module, a classification module, and a determination module; using the target data set to train the identification model; and after the model training is completed, processing target speech data by means of the identification model to obtain an identification result, wherein the identification result is used for indicating whether the target speech data is a synthesized speech.

A method for detecting aircraft flight conflict based on semantic parsing of control speech is provided. Firstly, the control speech is collected in real time, and the speech is converted into text through the control speech identification system combined with real-time radar data; inputting the identified control text into the control text intention identification model, further analyzing the intention, and judging whether the identified control text intention needs to change the flight state of the aircraft, if not, terminating the detection, and if necessary, extracting control instructions; extracting the key data needed for conflict detection from the identified control text; according to the key data and real-time radar data, performing the conflict detection algorithm to judge whether there is flight conflict; if it exists, a conflict alarm mechanism is triggered.

The invention relates to the technical field of natural language processing, and discloses a hatred speech recognition method and system.The system finely adjusts a plurality of models under various prompt configurations, hyper-parameters and input templates, reinforcement learning optimization based on human feedback is executed on the models, and finally reasoning results of the models are fused through a voting mechanism, so that a hatred speech recognition result is obtained. The final prediction output with higher stability and higher robustness is formed, and the adaptive capacity of the system under different input styles and expression modes is remarkably improved. According to the system, aiming at a hatred speech recognition problem in Chinese social media, fine-grained structure extraction, human feedback learning, reward modeling and reinforcement learning are organically combined for the first time, a hatred recognition task is divided into two stages of structure recognition and attitude judgment, and reward function modeling and near-end strategy optimization are introduced, so that the recognition accuracy of the hatred speech is improved. According to the method, the recognition capability of the model on the hidden hatred intention in complex speech can be remarkably improved.