Abstract: In one embodiment, a method includes receiving visual input of a media content a user is viewing from a client system associated with the user, determining one or more entities and one or more relationships between one or more of the entities associated with the media content based on the visual input, generating a personalized knowledge graph of the media content for the user based on the determined entities and relationships in real-time responsive to the visual input of the media content being received, generating a content summary based on the personalized knowledge graph, wherein the content summary corresponds to a current progress of the user within the media content the user is viewing, and sending instructions to the client system for presenting the content summary to the user.

Abstract: A user preference hierarchy is determined from user response to images. Images may be tagged using machine learning models trained to determine values for images. Products are clustered according to product vectors. Images of products within a cluster are clustered according to composition and groups of images are selected from image clusters for soliciting feedback regarding user preference for products of a cluster. Feedback is used to train a user preference model to estimate affinity for a product vector. A user may provide feedback regarding a price point and products are weighted according to a distribution about the price point. The distribution may be asymmetrical according to direction of movement of the price point. Filters may be dynamically defined and presented to a user based on popularity and frequency of occurrence of attribute-value pairs of search results and based on feedback regarding the search results.

Abstract: Systems, methods, and non-transitory computer-readable media can receive training data for training a machine learning model. The machine learning model is trained based on the training data. The machine learning model is trained to identify highlighted portions in audio content items. A first audio content item is received. One or more highlighted portions are identified in the first audio content item based on the machine learning model.

Abstract: There is provided a system configured to perform location-aware augmented reality tasks utilizing software and/or hardware components. For example, and not by limitation, a general embodiment may include an AR system. The system may include an ultra-wide band (UWB) communication system configured to estimate location information. The UWB system can include a memory and a processor which when executing instructions from the memory is configured to perform operations consistent with location-aware AR applications. For instance, the operations may include receiving a first UWB signal, the first UWB signal including information comprising fiducial data associated with a device transmitting the first UWB signal.

Abstract: In one embodiment, a method performed by a client device includes receiving, at the client device, an input of n-grams, and identifying media items from a media-item index. This media-item index may index media items stored by the client device and respective keywords associated with each media item, and each media item may be identified based on one or more of its associated keywords matching at least one of the n-grams of the input. An auto-suggestion with the identified media items may then be presented responsive to the input.

Abstract: Systems, methods, and apparatuses may provide optical lenses comprising liquid crystal metasurfaces. Systems and methods may include a lens system comprising a first optical lens and a liquid crystal metasurface formed on the first optical lens, and a pair of electrodes positioned on opposite sides of the first optical lens. The pair of electrodes may individually tune sections of the liquid crystal metasurface to adjust an optical characteristic of the optical signal. A waveguide may be configured to receive the adjusted optical signal passed through the first optical lens and may provide the optical signal to an image sensor for an image.

Abstract: To present one or more content items to users of an online system, the online system identifies a content evaluation pipeline including an order of a plurality of stages having one or more computer models for evaluating a likelihood of user interaction with a content item. The content evaluation pipeline selects a decreasing number of content items, from each stage of the order, according to the order of the stages. The online system optimizes the selection of content items selected at the plurality of stages of the content evaluation pipeline by training the computer models to predict content selection values that the subsequent model would generate for a content items in a training data set and content items that the subsequent model would select for input to the next stage of the content evaluation pipeline.

Abstract: In an aspect, first interaction data characterizing a first interaction of a user with a mobile device configured to operate an application client can be received. A plurality of segments of users of the application client can be determined, and each of the plurality of segments can be associated with one or more settings that characterize an operating parameter of the application client. The user can be assigned to a first segment of the determined plurality of segments based on a correspondence between a characteristic of the user and a characteristic associated with the first segment. The application client can be modified based on the one or more settings associated with the first segment. The modified application client can be provided for interaction by the user via the mobile device. Related systems, apparatus, techniques, and articles are also described.

Abstract: An online system for determining advertisements for retrieval is disclosed. The online system may assign a unique category path associated with a hierarchy of advertisement categories to a plurality of advertisements accessible by the online system to determine advertisement information features. The online system may determine a plurality of types of user engagements associated with prior user interactions by a user with one or more of the advertisements or types of advertisements associated with the advertisements to determine a set of user engagement features. The online system may select user engagement features, from the set, based on determined user engagement results to obtain user related features. The online system may associate the user related features with an assigned unique category path, associated with the ad information features, to generate inverted index features. The online system may apply the inverted index features as inputs to a neural network.

Abstract: A firmware massive update method using a flash memory includes: a firmware data registration step of receiving, from a manufacturer server, at least one of information of a user device that is a firmware update target, and firmware information and registering the received information as firmware data; a firmware data management step of receiving a request from a firmware update server in which the registered firmware data is stored, and storing and managing the registered firmware data in a specific area of a flash memory included in the user device via a network; and a firmware update execution step of executing a firmware update on the firmware data managed in the specific area of the flash memory included in the user device through the firmware update server.

Abstract: A computer-implemented method, comprising, by an artificial-reality (AR) design tool: receiving, through a user interface (UI) of the AR design tool, instructions to add a voice-command module to an AR effect, the voice-command module having an intent type and at least one slot, the slot associated with one or more entities; establishing, according to instructions received through the UI, a logical connection between the slot and a logic module configured to generate the AR effect depending on a runtime value associated with the slot; and generate, for the AR effect, an executable program configured to: determine that a detected utterance corresponds to the intent type and includes one or more words associated with the slot; select, based on the one or more words, one of the one or more entities as the runtime value for the slot; send the runtime value to the logic module according to the logical connection.

Abstract: A ticket exchange server is configured to determine a number of tickets to distribute for an event. The ticket exchange server accesses, for a stadium, training data describing attendance at historical events, historical opponents of a sports team, and a historical win/loss record of the sports team. The ticket exchange server trains a machine-learned model configured to predict an attendance for a future event at the stadium based on an opponent of the sports team at the future event and a current or predicted win/loss record of the sports team. The ticket exchange server selects an event for the sports team against an opponent and determines a predicted attendance using the machine-learned model. The ticket exchange server identifies a number of tickets greater than a capacity of the stadium to make available based on the predicted attendance and distributes the number of tickets to prospective attendees.

Abstract: A method may include (1) coating a segment of fiber optic cable with an adhesive substance, (2) forming a coil of the segment of fiber optic cable, (3) deforming the coil into a noncircular shape defining a slot external to the coil while obeying a minimum bend radius requirement for the segment of fiber optic cable, and (4) activating the adhesive substance to stabilize the noncircular shape of the coil. Various other methods and apparatuses, such as those for performing the deforming operation, are also disclosed.

Abstract: According to examples, an interconnect system for integrated circuits (ICs) may be fabricated by processing a substrate implanted with copper wells with a photoresist layer such that remaining portions of the photoresist layer expose portions of the copper wells; depositing a barrier layer over a top surface of the wafer, depositing a seed copper layer over the barrier layer; depositing a copper layer over the seed copper layer; planarizing the copper layer and portions of the barrier layer; depositing another copper layer over exposed portions of the substrate, the copper wells, and the interconnect cores; removing portions of the other copper layer between interconnects by processing the second copper layer with another photoresist layer; and removing remaining portions of the other photoresist layer on the interconnects.

Abstract: Methods, systems, and storage media for video navigation normalized by relevance are disclosed. Exemplary implementations may: obtain a plurality of videos recorded in a chronological order; stitch the plurality of videos together in a timeline based on the chronological order, the timeline comprising a visual menu; obtain information associated with the plurality of videos; determine, based on a measure of relevance, a ranking of videos among the plurality of videos; determine, based on the ranking of videos, one or more salient videos from among the plurality of videos, the one or more salient videos comprising a video that is within a range of the ranking of videos; obtain one or more representations associated with individual ones of the one or more salient videos; and cause display of the one or more representations in the timeline, the representations being arranged based at least in part on the chronological order.

Abstract: Various aspects describe an information platform for consistently integrating and/or quantifying the underlying principles of ESG into financial analyses, analytical tools, metrics, and/or available information on reviewed companies, business entities, etc. . . . , and further provide integration of analysis with community-based insight, contextual information and tools for readily understanding both. Various embodiments implement machine learning tools for curating data sources and incorporating the data sources into the knowledge platform. The incorporation of AI moderated information sources enables succinct views of often massive information pools, and further provides for transitions between types of information (e.g., qualitative, quantitative, and interactive data source (e.g., engagements, collaborative information, etc.)).

Abstract: A computer-implemented method for enabling shared media interactions may include (i) identifying a group message thread that is composed of messages between a group of users, (ii) detecting that a user in the group of users is interacting with a media item posted in the group message thread, (iii) in response to detecting that the user is interacting with the media item, initiating a shared media interaction session in the group message thread for the media item, and (iv) notifying the group of users that the shared media interaction session is available in the group message thread. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed flexible facial interface assemblies for head-mounted display systems may include a rigid support frame element and a flexible facial interface frame element. The rigid support frame element may be shaped and configured to physically support a display of a head-mounted display system in front of a user's eyes when the facial interface assembly is worn by the user. The flexible facial interface frame element may be configured to flex to conform to the user's facial features when the facial interface assembly is worn by the user. An outer periphery of the flexible facial interface frame element may be independently movable relative to an outer periphery of the rigid support frame element. Various other systems and methods are also disclosed.

Abstract: According to examples, a system for using artificial intelligence (AI) and machine learning (ML) techniques to generate and implement a configurable measurement platform is described. The system may include a processor and a memory storing instructions. The processor, when executing the instructions, may cause the system to access information associated with one or more events occurring on a platform with event activity, log and analyze the one or more events to generate measurement data associated with the event activity, and generate a metric associated with the measurement data. The processor, when executing the instructions, may then generate a computed metric value associated with the metric utilizing the measurement data, implement a platform computation utilizing the computed metric value, and facilitate a decision associated with the platform based on the platform computation.

Abstract: According to examples, a system for serving distributed inference deep learning (DL) models in serverless computing is described. The system may include a processor and a memory storing instructions. The processor, when executing the instructions, may cause the system to receive a request to initialize a container and request a first candidate server from an available resource finder and a second candidate server from a resource optimizer. The processor, when executing the instructions, may then implement the server allocator to prioritize use of one of the first candidate server and the second candidate server and provide feedback regarding the prioritized use of one of the first candidate server and the second candidate server.