Abstract: A generalized natural language—operating system (“GNL-OS”) for constraining a large language model (“LLM”) to provide deterministic output may be provided. The system may include delivery frameworks. Each delivery framework may deliver generalized natural language-program architecture natural language code (“GNL-PA NLC”) to the LLM. The LLM may operate on one or more hardware processors. The delivery frameworks may include a system framework. The system framework may deliver a system prompt structured in GNL-PA NLC code to the LLM. The system prompt may constrain the LLM to operate as an operative system. Upon receiving the system prompt, the LLM may operate as an operative system under one or more rules specified in the system prompt. The delivery frameworks may include a command framework. The command framework may deliver GNL-PA code to the LLM. The GNL-PA code may correspond to one or more commands executable within the operative system.

Abstract: A client application sends an application programming interface call to an interface of a content management system. The call specifies one or more content item search criteria. Sending the call causes the content management system to perform a lookup in a content item index to identify at least one content item that satisfies the one or more content item search criteria. Based on sending the call, the client application receives from the content management system a suggestion to attach the at least one content item to a text being displayed by at the computing system.

Abstract: A computer-implemented auditory biometric method generates a playlist file including at least one audio file. The method includes receiving user biometric data and applying user biometric data to a trained auditory biometric model to generate a playlist file. The auditory biometric model may be trained using training data including a plurality of historic records associated with a plurality of historic users. The method may include transmitting a playlist message including the playlist file to a user computer device for execution by the user computer device.

Abstract: A design image clustering method comprises the steps of: calculating the degree of similarity between pieces of target design element information selected by a user among pieces of design element information about a target design image and a plurality of reference design images stored in a database; clustering, on the basis of the degree of similarity, the reference design images into a plurality of clusters; and generating, for each cluster, at least one of at least one representative design image, an overlap image for the representative design image, and the degree of similarity between a center vector of each cluster and the target design image.

Abstract: An information processing apparatus includes a pose acquisition unit, a retrieval unit, and a display control unit. The pose acquisition unit acquires first pose information indicating a pose of a person shown in each of a plurality of reference images associated with a predetermined pose, and second pose information indicating a pose of a capture target person shown in a query image. The retrieval unit retrieves a reference image showing a person whose pose or action is similar to that of the capture target person shown in the query image, from among the plurality of reference images, based on a similarity degree between the first pose information and the second pose information. The display control unit causes a display unit to display at least one of the first pose information and the second pose information in a display mode according to an index used for retrieving the reference image.

Abstract: The present disclosure provides an information display method and apparatus, a medium, and an electronic device. The method includes: acquiring, in response to a search operation, a search result corresponding to the search operation, the search result includes a homepage of a target account that is associated with a search term corresponding to the search operation; displaying the homepage in a visual interface synchronously, and displaying an Easter egg superimposed on a picture that is displayed in a specified region of the homepage.

Abstract: The present disclosure provides a video search method, apparatus, server, and terminal device, the method comprising: acquiring search information; parsing the search information to acquire at least one field; when it is determined that there is a first video in a pre-built database that matches the search information according to the at least one field, acquiring the first video when a second video having an association with the search information is screened out from the database according to the first video, taking the first video and the second video as a search result and sending it to a terminal device, wherein the first video is a reply video which is generated by an object according to commentary content under any main video, and the second video comprises the main video, and/or, a video of the same category as the first video.

Abstract: The present disclosure relates to a method for creating a prompt, the method may include: receiving properties of a prompt; clustering the received properties into syntax elements; diversifying, by a large language model (LLM), each of the clustered properties into a plurality of properties; creating a prompt format database; selecting a list of prompt formats in the created prompt format database; and creating prompts by inserting each of the diversified properties, obtained from the diversifying of each of the clustered properties, into each prompt format in the list of prompt formats in accordance with the syntax element of the corresponding clustered property of the diversified properties. The present disclosure also relates to an apparatus for creating a prompt, and a computer program.

Abstract: A method of constructing a microbial identification database, the method comprising: (ST02) acquiring genome data for microorganisms from a genome database; (ST06) determining whether a criterion is satisfied by the genome data thus acquired; (ST16) for respective sets of the genome data that were determined that they satisfied the criterion, predicting proteins to be expressed; and (ST20A, 20C) constructing a mass-to-charge ratio database including mass-to-charge ratio lists, the mass-to-charge ratio lists being predicted for the respective sets of the genome data based on the proteins thus predicted.

Abstract: The present invention provides a system and method for curating media content. The method comprises: retrieving, by a computing device, from a database, a set of media content, and first set of metadata associated with the set of media content; sorting, by the computing device, the set of media content based on at least a first parameter, and at least a second parameter, wherein each of at least the first parameter and at least the second parameter comprises weightage values for each media content of the set of media content; and organizing, by the computing device, the set of media content based on a cumulative weightage value of each media content of the set of media content.

Abstract: A system may be configured to provide natural language search over security videos. In some aspects, the system may generate a first representation of sampled video information in a multidimensional format via a first machine learning model, and receive a request including a natural language input. Further, the system may generate a second representation of the natural language input in the multidimensional format via a second machine learning model that is a different from the first machine learning model, and determine that the first representation has a predefined relationship with the second representation. In addition, the system may present the second representation as a response to the request based on the first representation having the predefined relationship with the second representation.

Abstract: Described herein are systems and methods that highlight target objects in media content. The detection system trains a neural network to identify objects within an image. The detection system receives a text input requesting a search within the image and applies a search language model to the text input, which identifies a target object associated with the requested search. The detection system applies the neural network to the image to identify instances of the target object. The detection system modifies a user interface to include the image and modifies the image to highlight the identified instances of the target object. The detection system receives feedback that modifies the highlighted instances of the target object within the user interface and retrains the neural network based on the modified highlighted instances.

Abstract: An assigned subgroup that includes a plurality of entries is traversed by a prefetcher. It is determined that an expected number of entries associated with the assigned subgroup have been traversed. In response to determining that expected number of entries associated with the assigned subgroup have been traversed, it is determined that a last read entry associated with the assigned subgroup does not correspond to a last entry associated with the assigned subgroup. The prefetcher is preempted by stopping the prefetcher from obtaining a list of entries associated with a remaining portion of the assigned subgroup.

Abstract: A content collaboration platform, such as described herein, may include dynamic object tables that are configured to dynamically populate and/or update. Dynamic object tables may be configured to retrieve content (e.g., page content) from the content collaboration platform and automatically populate attributes from that content directly in the cells of the dynamic object tables. Updates to the page content may also be automatically updated within the table. In some examples, the dynamic object tables may be embedded within pages of the content collaboration platform and, within the same page, the user can edit, filter, and sort the dynamic object table without toggling to other tabs, applications, or documents.

Abstract: Recursive endorsements in a decentralized database minimize validation effort for a smart contract because there are fewer signatures to verify. Collected endorsements required by an endorsement policy are verified and signed by a set of trusted peers representing different organizations, the set of trusted peers being fewer than the number of endorsing peers. Prior to validation, the target transaction is submitted with all required endorsements and the signatures of the set of trusted peers. Validation is performed by verifying only the signatures of the set of trusted peers.

Abstract: Described are systems and processes for identifying relevant media content for users in response to natural language requests. The media content may be formed as a playlist of music. Requests may be sent to various models to obtain results, such as a large language model (LLM) and a local search model. Results may be selected from one or both models. The results may be used to fetch media content and provide user interfaces with a playlist with the media content to a user that submitted the request. User interfaces may include selectable search terms suggested for the user and animations during processing of the request.

Abstract: Methods and systems described herein can provide a computing device configured to display on a display screen of the computing device a first window listing a set of entries and a second window that can add entries directly into the set of entries and that can search for entries having selected features. The second window can dynamically display one or more keywords related to an entry as the entry is being created. The keywords can be determined as the entry is created and can be associated with the entry as the entry is added to the set of entries. The second window can receive a search request with the keywords and in response, cause the entries associated with the keywords to be displayed in the first window.

Abstract: Devices, systems, and methods for machine learning-based micro-category generation may include a method including identifying, for a respective cluster of content titles or items, a top-K most representative group of the content titles or items; generating, by a machine learning model, based on a first input prompt, the common theme of the top-K most representative group; generating, by the machine learning model, based on the second input prompt and filtering, a group of content titles or items matching the common theme; generating, by the machine learning model, based on a third input prompt, a name for the group of content titles or items matching the common theme; and presenting the group of content titles or items matching the common theme and the name for the group of content titles or items matching the common theme via a user interface of a streaming media application.

Abstract: A computer system is provided that is programmed to select feature sets from a large number of features. Features for a set are selected based on metagradient information returned from a machine learning process that has been performed on an earlier selected feature set. The process can iterate until a selected feature set converges or otherwise meets or exceeds a given threshold.

Abstract: A device may receive video data and description data, and may extract image embeddings and video embeddings from the video data. The device may receive a query that includes a natural language description of a scene to be retrieved from the video data. The device may split the query into a first subquery and a second subquery, and may process the first and second subqueries, with a model, to generate a first query embedding and a second query embedding. The device may generate appearance similarity measures based on the image embeddings and the first query embedding, and may generate dynamic similarity measures based on the video embeddings and the second query embedding. The device may combine the appearance similarity measures and the dynamics similarity measures into combined similarity measures, and may retrieve video segments based on the combined similarity measures.