Abstract: A method for identifying cyberbullying may include obtaining content from a first electronic device associated with a first user; identifying, using a machine learning model, the cyberbullying of the first user based on the content obtained from the first electronic device; and providing anonymized information related to the cyberbullying to a second electronic device associated with a second user.

Abstract: A multimodal embedding modifier generates a modified seed search selection embedding for providing a set of search results. The multimodal embedding modifier enhances the ability and accuracy of identifying a user's true intent when searching the online marketplace. For example, embodiments disclosed herein can allow a user to navigate multiple modalities for an item. In some embodiments, a user may select a search result corresponding to an initial search query, and further modify the selected search result by inputting a modifier (e.g., a textual modifier). The multimodal embedding modifier can be trained using a training dataset including a text embedding, an image embedding, another type of embedding, or a combination thereof.

Abstract: Search queries are received and search results are provided. Interaction tracking is used to determine with which search results users interact. The search results having received interactions can be represented as item vectors, which can include a vector representation of a portion of the search result, such as a title, description, or image. For each search query, the item vectors are aggregated, such as by averaging the item vectors. The search queries are stored in an item dataset as collected search queries respectively associated with the aggregate item vectors. When a new search query is received, a search query vector can be compared to the aggregate item description vectors to identify collected search queries that are related. The related collected search queries can be provided as search query recommendations or search results associated with the collected search queries can be provided in response to receiving the new search query.

Abstract: One embodiment comprises a data processing system for populating selections in an evaluation operator interface. The system may record voice sessions of phone calls, transcribe the voice sessions and store transactions including the voice sessions and transcripts. The system receives a request from a client computer for an evaluation to evaluate a transaction that was assigned an automated score according to the automated scoring template based on a transcript of the transaction having matched a lexicon. The system generates the evaluation. Generating the evaluation comprises setting an answer control for a question in the evaluation to a preselected answer based on a defined correspondence between the automated score and the preselected answer, the preselected answer selected from a defined set of acceptable answers to the question. The system may generate page code for the answer control that sets the answer control to the preselected answer.

Abstract: Systems and methods are disclosed to improve disaster recovery by implementing a scalable low-loss disaster recovery for a data store. The disaster recovery system enables disaster recovery for a linearizable (e.g., externally consistent) distributed data store. The disaster recovery system also provides for a small lag on the backup site relative to the primary site, thereby reducing the data loss by providing a smaller data loss window compared to traditional disaster recovery techniques. The disaster recovery system implements a timestamp for log records based on a globally synchronized clock. The disaster recovery system also implements a watermark service that updates a global watermark timestamp that a backup node uses to apply log records.

Abstract: A system and method for creating avionic databases is disclosed. The system may receive an avionic database of an area, surveyed coordinate points of the area, and new satellite imagery of the area. The system may georeference the new satellite imagery using the surveyed coordinate points as ground control points, generating georeferenced coordinate points of the new satellite imagery in different locations of the area. The system may create a second avionic database based on the georeferenced coordinate points.

Abstract: Proposed are a blockchain-based high-speed data processing device and a method thereof, which can manage data in a data structure format that can maximize performance when a blockchain system processes data. According to an embodiment, when <Key>:<Value> is stored, the final value is stored in a Log Structured Merge (LSM) tree data structure format, and the key used for storing the final value is separately stored in a B+ Tree data structure format, so that storage performance and reference (read) performance can be improved compared to those of conventional methods.

Abstract: A set of queries from an application executing on a client computing device is obtained. A first database based on the set of queries is searched to select a set of event types. A set of predicted parameters associated with the set of event types is sent to the application. The application includes instructions to obtain a first parameter and the set of predicted parameters via a user interface of the application and generate a message comprising the first parameter and an indicator identifying the set of predicted parameters. The first parameter and the indicator are obtained via the second message. A combined query including the first parameter and the set of predicted parameters is generated in response to obtaining the indicator. A vehicle record from the vehicle database is obtained based on the combined query. Values of the vehicle record are sent to the client computing device.

Abstract: Extracting key-value attributes from unstructured data includes a parser receiving a first incident comprising key-value attributes and a second incident comprising key-value attributes, and parsing the first incident for the key-value attributes and the second incident for key-value attributes. A machine learning (ML) model perform a pairwise comparison of one or more common key-value attributes associated with the first incident and the second incident. A cosine similarity module computes a cosine similarity between the one or more common key-value attributes to generate a score for each of the one or more common key-value attributes associated with the first incident and the second incident, and generates a final score between the first incident and the second incident by averaging all cosine similarity scores computed for each of the one or more common key-value attributes.

Abstract: A visual search recommendation engine utilizes an image quality indication model for a visual search recommendation. Specifically, the visual search recommendation engine receives a search image as a search query at a search engine. The visual search recommendation engine provides the search image as an input into the image quality indication model, which is trained to output an image quality indication based on image aspects of the search image. A plurality of images are identified from an image corpus. The visual search recommendation engine determines an image similarity based on a comparison between the plurality of images from the image corpus and the search image. The image quality indication and the image similarity indicate a search query performance for the search image. A first image exceeding the search query performance is identified. The first image is provided for display at the search engine.

Abstract: Methods and systems disclosed herein may optimize write operations in a transaction to reduce the number of operations to the point where each of the write operations may be performed in parallel. The writer optimizer may review a first write operation and a second write operation to determine whether the first write operation and the second write operation write to the same row in the same table. When the first write operation and the second write operation write to the same row in the same table, the first write operation and the second write operation may be combined to generate a functionally-equivalent third write operation. The third write operation may reduce the number of operations in the transaction by replacing both the first write operation and the second write operation.

Abstract: A computer implemented method includes obtaining multiple configuration files that include configuration commit histories, detecting patterns in parameter values in the configuration files to generate file-based rules for configuration parameters, detecting patterns in parameter values in the configuration files to generate history-based rules using commit histories for the configuration parameters, and exposing the rules to calling programs.

Abstract: A system and method are presented that improves search results based on skills associated with individuals. The system monitors an individual's contributions to a third-party forum and associates the contributions with skills. The system also monitors learnings received by the individual. Assignments are made to associate the individual with the skills and to assign points. Users request a search based on search skill parameters, and the system uses forum-based data and non-forum data to identify and sort individuals. Skill and point assignments are made temporarily, and assignments are deleted upon the expiration of a time-to-live clock. Save events can permanently save data. Local forum data is also stored and can be used to restore deleted assignments. Points are used to improve avatars used when interacting with searching users.

Abstract: Systems and methods for debiasing a recommendation engine are disclosed herein. A search query associated with a user profile is received at a recommendation engine. Control circuitry generates a result set of items of content based on the search query and generates a bias score for a content attribute based on the result set. The control circuitry also generates a time-averaged bias score for the content attribute based on a plurality of search queries associated with the user profile. Based on the bias score and the time-averaged bias score, the control circuitry determines whether a bias is signaled for the content attribute. Finally, the control circuitry outputs, for display via a computing device, the result set or a debiased result set based on a result of the determination of whether the bias is signaled.

Abstract: Implementations of the disclosed technologies pre-fetch search results. Implementations receive first input from a search session of a user device, where the first input includes at least a portion of a search term but does not initiate a search. Implementations determine context data associated with the first input, determine that a combination of the first input and the context data satisfies a pre-fetch threshold, determine intent data based on at least a portion of the context data, generate a search query based on the first input and the intent data, and pre-fetch a first subset of search results based on the search query. In response to a second input received subsequent to the first input, where the second input contains an initiate search signal, implementations initiate rendering of the pre-fetched first subset of search results in the search session at the user device.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for classifying data objects. One of the methods includes obtaining data that associates each term in a vocabulary of terms with a respective high-dimensional representation of the term; obtaining classification data for a data object, wherein the classification data includes a respective score for each of a plurality of categories, and wherein each of the categories is associated with a respective category label; computing an aggregate high-dimensional representation for the data object from high-dimensional representations for the category labels associated with the categories and the respective scores; identifying a first term in the vocabulary of terms having a high-dimensional representation that is closest to the aggregate high-dimensional representation; and selecting the first term as a category label for the data object.

Abstract: A popularity estimation device includes: an acquisition unit configured to acquire a first map image including a first particular point and to acquire a plurality of second map images including a second particular point which is a particular point; a generation unit configured to generate a first feature vector and a second feature vector by inputting the first map image and the second map images to a geographical feature generation model with a map image as an input and with a feature vector indicating geographical features of the map image as an output; a score calculating unit configured to calculate a popularity score of the first particular point from a popularity score of the second particular point based on a degree of similarity between the first feature vector and the second feature vector; and an output unit configured to output the popularity score of the first particular point.

Abstract: A system and method for a patent search and analytics software tool that finds prior art for each claim limitation/element by breaking up claims into individual claim limitations. Once the claims are separated, the system finds the best prior art for each of the individual separate, different claim limitations/elements. Additionally, the software finds the best prior art for entire claims, including non-patent literature (NPL) searching. The system takes into account the limitations of the claim under consideration (query claim of query patent), the text of the art, the link structure of the citation network, and the patent classification and then constructs a network that consists of two types of nodes: (i) the art (patents and non-patent literature) and the (ii) classes of the patent classification. Each art node is linked to all the art nodes that it cites and is linked to all the classification nodes that it belongs to.

Abstract: Techniques for generated regular expressions are disclosed. In some embodiments, a regular expression generator may receive input data comprising one or more character sequences. The regular expression generator may convert character sequences into a sets of regular expression codes and/or span data structures. The regular expression generator may identify a longest common subsequence shared by the sets of regular expression codes and/or spans, and may generate a regular expression based upon the longest common subsequence. A negative example may be used to generate the regular expression. Context from the negative example may be determined in order to generate the regular expression.

Abstract: The present disclosure relates to methods and systems for controlling content distribution. The methods and systems receive an original search query and expand the original search query by adding one or more neutral words to create an expanded search query. The methods and systems automatically classify the original search query using the expanded search query in an accurate manner and use the classification to control content distribution.