759results about "Metadata audio data retrieval" patented technology

The invention provides a multi-modal-based data retrieval enhancement method, which solves the problems of semantic gaps and outdated information among different modal data through cross-modal semantic alignment and real-time data source integration. And secondly, by constructing a multi-modal feature space and introducing a dynamic retrieval layer, the retrieval speed and accuracy are improved. In addition, by combining sparse retrieval and dense retrieval methods, the accuracy and the recall rate are balanced, and the retrieval result is optimized. And finally, through the steps of post-processing and result optimization, the correlation of the retrieval result and the user experience are further improved. The method is not only suitable for data retrieval of common modes such as texts, images and videos, but also can be expanded to any multi-mode scene, and provides powerful technical support for cross-mode retrieval.

A scoring system and method identifies personal attacks in a piece of audio content and generates a civility score for the piece of audio content that can differentiate between personal attacks and vernacular / casual banter. The piece of audio content may be a podcast.

The invention provides an acoustic data storage and retrieval method combined with computer cloud computing, which comprises the following steps of: performing time sequence division processing on original acoustic data to generate an audio frame sequence with a timestamp mark; based on a context window of the audio frame sequence, acoustic features are extracted, and a context-aware acoustic feature vector is constructed; constructing a time sequence sensing distributed index structure according to the acoustic feature vector, and dynamically storing the streaming acoustic data to obtain the time sequence sensing distributed index structure; and executing cross-modal similarity retrieval on the query audio and the distributed index structure to obtain a target fragment sequence in space-time association with the query audio. By adopting the method, the retrieval precision and efficiency of the acoustic data in a complex time sequence association scene can be improved through a time sequence perception storage architecture and a cross-modal retrieval mechanism.

A method, a device and / or a system of a generative AI computing platform implemented using a processor communicatively coupled to a memory is disclosed. In accordance therewith, context relevant to an interaction of a user with the generative AI computing platform is determined from an input thereto indicative of the interaction of the user. In addition to the context, a current location of the user is determined from the input, data associated with the input and / or metadata of the user outside the input and the data associated with the input. Also, one or more operation(s) contextually relevant to and reflective of the determined context and the determined current location of the user is automatically executed as associated with a response to the interaction of the user.

The invention relates to the technical field of multimedia, and provides a song recommendation method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring song-based real-time user interaction data and historical user use data; a target model is created based on the real-time user interaction data and the historical user usage data, and the target model is a model representing user preferences; a target music feature and a candidate song set similar to the currently played song are obtained, and the target music feature is a feature obtained after multi-modal music features are fused; and creating a dynamic queue of the candidate song set based on the target model and the target music features to obtain a to-be-played song set. According to the method, diversification of data analysis factors of song recommendation is enhanced, intelligent dynamic adjustment of the music playing queue is realized, highly personalized and dynamically responsive music playing experience is provided for the user, the accuracy of the played songs is improved, and thus the personalized song listening experience of the user is improved.

An electronic device associated with a media-providing service stores, in a vector space, a plurality of respective vector representations for respective media content items. The electronic device receives a user input, including a text string. The electronic device generates, using a neural network, a structured query based on the text string. The electronic device determines, based on the structured query, whether to generate a vector representation of a portion of the text string. When the electronic device determines to generate the vector representation of the portion of the text string, it generates the vector representation of the portion of the text string, wherein the vector representation is embedded in the vector space, and identifies a set of media items using the vector representation of the portion of the text string. And the electronic device provides one or more select media items from the set of media items to a user.

The invention relates to a voice interaction method based on a dynamic tag library and terminal equipment. The method comprises the following steps: receiving voice interaction information sent by a user; performing intention recognition on the voice interaction information based on a preset dynamic tag library, and determining an intention recognition result matched with the voice interaction information; wherein the preset dynamic label library comprises a user behavior label library and a dynamic content label library; the user behavior tag library comprises behavior tags related to user preferences and behavior habits; the dynamic content label library comprises dynamic content labels related to user preferences; based on the intention recognition result, executing an action sequence corresponding to the intention recognition result; the interaction intention of the user can be more comprehensively and accurately understood, so that accurate interaction is realized, and the individuation degree and accuracy of voice interaction are improved.

The invention discloses an information retrieval method, a related device, equipment and a storage medium, and is applied to the field of artificial intelligence. The method comprises the steps of obtaining to-be-queried information and target instruction information; based on the to-be-queried information and the target instruction information, obtaining a first token sequence through a feature extraction layer included in an information retrieval model; performing bidirectional attention coding on the first token sequence through a backbone network included in the information retrieval model to obtain a second token sequence; pooling the second token sequence to obtain a target embedded vector; and according to the target embedding vector, obtaining a preset number of candidate information with the highest similarity as a retrieval result. According to the method, the complexity and the calculation cost of model deployment are reduced, and the target embedding vector generated by the model is more discriminative, so that the accuracy of a retrieval result is improved.

The invention relates to the technical field of music score matching, in particular to an audio and music score matching method and system, and the method comprises the steps: providing a reference database of a music score, which comprises multiple segments of expected playing characteristics of a plurality of music units; acquiring first note data generated by playing of the user at first time; judging whether second note data generated at adjacent second time before the first time is successfully matched with the first target music unit or not; if yes, determining a first dynamic prediction sequence at the first time according to the first target music unit; judging whether the corresponding music unit obtained in the first dynamic prediction sequence is a second target music unit of the first note data or not; and if yes, marking the second target music unit as the current playing node. According to the music score matching technology, the matching time axis can be dynamically adjusted in combination with the playing state, and then positioning matching of the audio and the music score can be rapidly achieved through the limited matching calculation amount.

The invention discloses an interface display method and device in a music application program, equipment and a storage medium, and relates to the technical field of computers. The method comprises the following steps: displaying a vertical-screen song playing interface, wherein vertical-screen song playing information of a first song is displayed in the vertical-screen song playing interface; in response to a switching operation for the song playing interface, switching the song playing interface in a vertical screen form into a song playing interface in a horizontal screen form; the song playing information of the first song in the horizontal screen form is displayed in the song playing interface in the horizontal screen form, and the song playing information comprises at least one of a song cover image, song basic information, a song function control and a song background image. According to the method, the song playing information in the landscape form is displayed, so that the song playing information is not limited by the screen width any more, the picture ratio of the basic song information is increased, and the display degree of the basic song information is improved.

The invention discloses an intelligent control method and device for a die casting deburring process. The intelligent control method comprises the steps that a grinding head advancing track is generated according to a trimming contour of a die casting; the method comprises the following steps: extracting burr characteristics by analyzing a die casting image, and matching proper deburring audio data according to the burr characteristics; the abrasion degree of the grinding head is detected, the advancing track of the grinding head is corrected, and meanwhile corresponding grinding head control data are generated; and in the deburring process, audio feature data are collected in real time and compared with preset audio data, and if the similarity does not reach the standard, the grinding head track and the control data are adjusted. By accurately analyzing the abrasion of the grinding head and automatically correcting the track and control parameters of the grinding head, the high-quality and high-efficiency deburring effect is ensured in combination with real-time audio monitoring, manual intervention is reduced, the service life of equipment is prolonged, and the product consistency is improved.

The invention relates to an interactive music intelligent generation method and system adaptive to a scene space, and the method comprises the steps: building a conditional diffusion model frame through sensing the behavior, emotional state and space environment change of a user, generating music matched with a current scene in real time, achieving the deep fusion of the music and user experience, and improving the user experience. And the generated music is dynamically adjusted to ensure that the generated music not only can flexibly cope with scene changes, but also can meet personalized requirements of the user, so that the interaction experience of the user with the music generation system in the scene is improved, and the diversity and intelligence of scene scheme generation are improved.

This application relates to a method for matching music with a video performed by a computer device, and a storage medium. The method includes: determining a cut speed of a video; determining a long-time audio speed corresponding to each of a plurality of pieces of candidate music according to a high-scale point and a music duration of the candidate music; selecting matched music from the pieces of candidate music according to the cut speed and the corresponding long-time audio speeds; determining, according to a video duration of the video and a high-scale point corresponding to the matched music, a short-time audio speed corresponding to each music clip in the matched music; and determining a target music clip in the matched music according to the cut speed of the video and the corresponding short-time audio speed, and synthesizing the target music clip and the video to obtain a target video.

The invention relates to the technical field of deep learning, and discloses a video music recommendation method and device based on deep learning, equipment and a medium, and the method comprises the steps: obtaining a video to be matched with music and background music to be matched with the video to be matched with music; extracting key frame data of the video to be dubbed with music in each time period, and calculating a change value of the key frame data of the video to be dubbed with music in each time period; extracting feature information of the background music, and calculating a melody value of the background music according to the extracted feature information of the background music; and matching the change value of the key frame data of the video to be dubbed with the melody value of the background music, and recommending the background music to the video to be dubbed according to a matching result. According to the video music recommendation method based on deep learning, the efficiency of recommending background music to the video to be matched with music is effectively improved, and a large amount of labor cost is saved.

The system receives a voice query at an audio interface and converts the voice query to text. The system can determine pronunciation information during conversion and generate metadata that indicates a pronunciation of one or more words of the query, include phonetic information in the text query, or both. A query includes one or more entities that may be more accurately identified based on pronunciation. The system searches for information, content, or both among one or more databases based on the generated text query, pronunciation information, user profile information, search histories or trends, and optionally other information. The system identifies one or more entities or content items that match the text query, and retrieves the identified information to provide to the user.

The invention discloses a song teaching template generation method and system, equipment and a program product. The method comprises the following steps: acquiring song audio information of a target song audio; wherein the song audio information comprises lyrics, lyric timestamps, time points corresponding to paragraphs and audio representation information; performing various intra-sentence teaching feature analysis on the lyric timestamps and the audio representation information of the target song audio, and marking various analyzed intra-sentence teaching features on lyrics; wherein the in-sentence teaching features comprise at least one of accent, treble, long tone, tremolo, gliding and breathing points; the singing difficulty feature comprises at least one of high pitch and long pitch; the skill features comprise at least one of tremolo and glide; on the basis of the time point corresponding to each paragraph and the audio representation information, performing in-segment teaching feature marking on the loudness gradient part of each paragraph in the lyrics of the target song audio; and generating a song teaching template based on the marked lyrics of the target song audio and outputting the song teaching template.

The invention relates to the technical field of routing inspection information data processing, and discloses a writing type routing inspection information input data processing system and method based on voice recognition, and the system comprises a routing inspection information input module, a routing inspection information modification module, and a routing inspection information modification execution module. Based on scientifically preset audio parameters of different dialect accent types, in combination with an intelligent audio matching algorithm and the audio parameters of the accent of the worker, accurate analysis of the accent type of the worker is carried out, and self-adaptive voice recognition based on the accent type of the worker is realized to input inspection information; the accent type parameters of the working personnel are combined with an intelligent recognition algorithm and different dialect accent type voice recognition parameters to perform customized matching on the types of personal voice recognition data packets of the working personnel and accurately acquire inspection information, so that the inspection information is intelligently, accurately and efficiently acquired based on accent characteristics of different working personnel; and the convenience and applicability of the inspection information entry system are improved.

The invention provides a maintenance guidance method and device for a household electrical appliance. The method comprises the following steps: constructing a unified vector database in advance based on text data, image data and audio data related to a household electrical appliance fault; acquiring field condition information of the household electrical appliance, wherein the field condition information comprises video information, audio information and / or text information; respectively carrying out image matching, audio matching and / or text matching on the field condition information based on the unified vector database, and outputting knowledge fragments as retrieval results; and inputting a retrieval result into the large language model, and outputting a maintenance guidance scheme of the household electrical appliance. According to the method, a retrieval and reasoning framework based on multi-modal feature direct fusion can be constructed, information dimension reduction loss is avoided by directly analyzing image, audio and text features of faults, and the comprehensive accuracy of fault diagnosis is improved to 95% or above; and through a dynamic knowledge retrieval and enhanced generation technology, the risk of'knowledge illusion 'possibly occurring in the professional field of the general large language model is effectively reduced.

A method of generating congruous metadata is provided. The method includes receiving a similarity measure between at least two multimedia objects. Each multimedia object has associated metadata. If the at least two multimedia objects are similar based on the similarity measure and a similarity threshold, the associated metadata of each of the multimedia objects are compared. Then, based on the comparison of the associated metadata of each of the at least two multimedia objects, the method further includes generating congruous metadata. Metadata may be tags, for example.

The invention discloses a music education practice system and a method thereof, and relates to the technical field of education. The system comprises a music score library, a teacher unit, a student unit and a display unit. Wherein the music score library is used for storing online open source music scores; the teacher unit is used for carrying out music score analysis on the music score and splitting the music score into corresponding signal components according to a singing mode and a playing mode to form an original audio; the student unit is used for practicing target music through peripheral equipment and recording by taking the length of each piece of music as a cycle to obtain a practicing audio; and the display unit is used for dynamically displaying the original audio and the practice audio in the form of visual notes, and deducting scores when the original audio and the practice audio have signal inconformity to obtain a final practice scoring result. According to the invention, the user can be assisted to improve the convenience of music practice.

The invention discloses a music generation method based on multi-modal cross attention and dynamic feature fusion, which is characterized by comprising the following steps of: extracting semantic features of a text by using a pre-training language model, and extracting image region features through a pre-training visual model; a multi-head cross attention module is used, text features are used as Query, image features are used as Key and Value, and a guide module pays attention to features, strongly related to texts, in pictures; constructing a small neural network prediction modal weight, and dynamically weighting and fusing text features and picture features; and extracting emotion information in the text and the picture by using a long-short-term memory network and a convolutional neural network, and guiding a diffusion model to generate symbol music according to the fusion features. Compared with the prior art, the method has the advantages of efficiently capturing potential information of texts and pictures, being high in cross-modal information fusion capability, enhancing the consistency of potential emotions of music and input information and the like, can improve the matching degree of music and input to a certain extent, and has a good application prospect.

The present disclosure relates to a system and method for selecting and generating audio using a large language model. The method includes receiving from a user a text-based prompt for a desired song, generating a song specification from a prompt that includes the text-based prompt and instructions on how to create a suitable instruction file format for representing the requested song, for each of the list of tracks in the song specification, generating a ranked list of potential sound loops matching the song specification for a selected track, selecting a sound loop from the ranked list of potential sound loops for each of the list of tracks, and generating a track specification file including the sound loop selected for each of the list of tracks.

A system and method for media content management include creating, via a digital vault, a container file comprising media content submitted by a first user and content metadata; verifying, via the digital vault, a completeness of the content metadata associated with the media content in the container file; classifying, via the digital vault, the container file based on the completeness of the media content; capturing, via the digital vault, event metadata when a second user gains access to the container file, the event metadata comprising at least one of identification of the second user, an activation timestamp, a duration of access, portions of the container file accessed, and changes to the container file; and enabling a private communication channel between parties affiliated with the media content to permit messaging among the parties affiliated with the media content via the private communication channel.

The invention relates to the technical field of Internet information processing, in particular to a sound effect recommendation method, system and device based on user preference and scene awareness, and the method comprises the steps: carrying out the deep analysis of collected user preference data, and constructing a user preference model; when an audio recommendation request of a user side is received, scene perception information matched with a current vehicle position is obtained, and a current scene is determined; matching the current scene with the user preference model to obtain a candidate sound effect set matched with the current scene and the user preference; checking each candidate sound effect data in the candidate sound effect set, screening the candidate sound effect data according to the sound effect data, and generating a recommendation result; and feeding back the recommended sound effect contained in the sound effect recommendation result. According to the scheme, the personalized requirements of the user can be fully considered, and the preferable sound effect recommendation information is provided for the user, so that the satisfaction degree of the user on sound effect recommendation is improved.

The invention relates to the technical field of audio playing, and discloses an audio playing method and device capable of keeping contexts continuous, equipment and a storage medium. Matching an audio context model conforming to a preset rule from a playing record database; analyzing the audio context model, and extracting audio playing state information and sound effect configuration information recorded in the audio context model; and adjusting the sound effect of the target playing device, and controlling audio playing according to the audio playing state information. Through the playing control mode, switching playing of multiple sound sources can be achieved, continuous playing after switching can also be achieved, and the problems that audio playing repetition is high after traditional switching, and the user experience feeling is reduced are solved.

The invention relates to the technical field of data processing, and discloses a recommended song menu customization method, device and equipment, a storage medium and a computer program product, and the method comprises the steps: obtaining user identity information of a current user, obtaining a music playing record and a user portrait of the current user based on the user identity information, and obtaining a recommended song menu based on the music playing record; and obtaining a user preference song list of the current user, and screening songs in the user preference song list according to the user portrait to obtain a recommended song list. Therefore, the music playing record and the user portrait of the corresponding current user can be obtained according to the obtained user identity information, the user preference song list of the current user is obtained based on the music playing record of the user, and the songs in the user preference song list are screened according to the user portrait. Due to the fact that the music playing record of the current user and the user portrait are combined, the effect of more accurately matching the song listening requirement of the user is achieved.

The present application discloses a method, device and storage medium for processing meeting minutes, relates to the field of voice conference technology, and discloses a method for processing meeting minutes, including: sending a meeting request to a server to create a meeting, collecting audio frames through a conference recording component and sending them to the server; receiving meeting minutes information obtained by the server based on the audio frames, ASR technology and a pre-trained large language model; obtaining a user's meeting query request; displaying matching content of the meeting minutes information according to the meeting query request; realizing real-time display of meeting content, being able to summarize the meeting content from a semantic level, and timely pushing meeting completion information, so as to facilitate users to understand the meeting content in real time and efficiently.

Techniques for accessing shared content based on a recipient's subscriptions are described. A recipient user may receive a notification that a sender user shared content. The recipient user may request to output the shared content. The system may determine which content providers the recipient user is subscribed to, and may perform entity resolution processing to identify potential content, from the subscribed content providers, corresponding to the shared content. The system may output the identified potential content in response to the request to output the shared content.

A system includes a computing platform having processing hardware, and a memory storing software code and a machine learning (ML) model-based feature classifier. When executed, the software code receives media content including a first media component corresponding to a first media mode and a second media component corresponding to a second media mode, encodes the first media component using a first encoder to generate multiple first embedding vectors, and encodes the second media component using a second encoder to generate multiple second embedding vectors. The software code further combines the first embedding vectors and the second embedding vectors to provide an input data structure for a neural network mixer, process, using the neural network mixer, the input data structure to provide feature data corresponding to a feature of the media content, and predict, using the ML model-based feature classifier and the feature data, a classification of the feature.