Abstract: In embodiments of the present disclosure, there is provided a method for generating a stylized icon automatically. After a query text inputted by a user is obtained, a trained generator is used to generate a structured icon that can characterize a structure of an object, and then the structured icon is stylized, such as performing color padding or adding other styles, so as to generate a high-quality stylized icon for the user. In embodiments of the present disclosure, a structured icon and a stylized icon are generated respectively at two stages, where the structured icon can clearly characterize the structure of the object, while the stylized icon can be richer in color and style. Therefore, the stylized icon generated according to embodiments of the present disclosure has a higher quality and is more realistic, thereby improving the user experience of icon generation.

Abstract: Technologies are disclosed that enable a computing system to present a structured arrangement for tracking content items on a shared user interface (UI) during a communication session. The structured arrangement is a list that is displayed in a specific region of the shared UI. Inclusion of content items in the list makes it easier for users to locate and interact with those content items throughout the communication session. The ability to create and manipulate the list may be limited to only certain users such as a moderator. Use of this list can promote inclusivity and fairness. For instance, inclusion in the list may prevent content items from being forgotten or ignored. Additionally, the names of users who contributed the content items may be shown in the list thereby providing recognition for those users.

Abstract: A technique partitions a user's original query into plural smaller component queries, each of which has a common part and an instance-specific part. The technique distributes the component queries to plural processor instances of a processor. The plural processor instances transform the respective component queries into query-component responses by acting in parallel, independent of each other. The technique generates a final response based on the query-component responses, e.g., by assembling the component-query responses into the final response. The technique reduces latency because the processor instances work on parts of the user's original query at the same time, rather than as a single stream of consecutive tokens. The plural processor instances have access to a shared cache memory, and utilize relevant data that has been computed in response to previous queries.

Abstract: A classifier may be trained with less than all datasets manually annotated with labels. A small subset of verbatims may be manually labeled with topic labels as seeds. Data augmentations can be used to acquire seed verbatim sets for known topics and to assign temporary pseudo labels to the rest of the verbatims based on their vector space proximity to the labeled seed verbatims. The training may involve classification epochs during which embeddings are updated with the assumption that the pseudo labels are ground-truth labels. The training may also involve labeling epochs during which the updated embeddings are used to update the vectors corresponding to the verbatims, and pseudo labels are updated based on updated vector coordinates in the vector space. As the training process progresses through the epochs, the embeddings will converge.

Abstract: According to implementations of the subject matter described herein, there is provided a solution for secure execution of a machine learning network. An operation of a first network layer of a machine learning network is executed in an uTEE of a computing device based on an input of the first network layer and a first set of modified parameter values, to obtain a first error intermediate In output. The modified parameter values are determined by modifying at least one subset of parameter values of the first network layer with first secret data. A first corrected intermediate output is determined in a TEE of the computing device by modifying the first error intermediate output at least based on the input and first secret data. A network output is determined based on the first corrected intermediate output. In this way, it is possible to protect the confidentiality of the machine learning network.

Abstract: Solutions for enforcing both syntactic and semantic correctness of domain-specific data queries include: receiving a data query; constraining an operation of the data query to enforce semantic correctness, wherein enforcing semantic correctness comprises: determining semantic information associated with each of a plurality of data entities; based on at least the semantic information, determining whether combining the two data entities of the plurality of data entities is allowed or is not allowed; based on at least determining that combining the two data entities is allowed, permitting combining the two data entities in the data query; and based on at least determining that combining the two data entities is not allowed, preventing combining the two data entities in the data query; and executing the data query. Some examples further include basing the permission on (at least) an ontology between the semantic information of the two data entities and a set of axioms.

Abstract: Disclosed in some examples are methods, systems, and machine-readable mediums in which interference is mitigated on a wireless radio frequency channel used by multiple computing devices by adjusting (e.g., lowering) the power level of one or more computing devices (e.g., mobile devices such as User Equipment (UE) devices) and using successive interference cancellation to decode the combined signals. Successive interference cancellation may first decode a strongest signal of a combined signal of all computing devices. The decoded signal is then subtracted from the combined signal. The strongest signal of that combined signal is then decoded and then subtracted and so on until all signals are decoded.

Abstract: Some embodiments bridge a gap between focusing on security alerts raised by conditions and events that have already occurred, and focusing on vulnerabilities that might be exploited in the future. Alerts are organized into alert categories, vulnerabilities are organized into vulnerability categories, and are optionally supplemented with misconfiguration categories. Correlations are identified between alert categories and vulnerability or misconfiguration categories, and the correlation values noted, to produce category association rules. The alerts, vulnerabilities, and other security findings are gathered in some situations from multiple similar environments, and in some cases are filtered to pertain to similar resources or similar configurations. The category association rules are utilized to perform cybersecurity prioritizations such as assigning priority levels to alerts and assigning likelihood levels to potential breaches.

Abstract: Trap circuits for use with superconducting integrated circuits having differential capacitively-coupled resonant clock networks are described. An example superconducting integrated circuit (IC) includes a first superconducting circuit comprising: (1) a first Josephson junction (JJ) coupled via a first capacitor to a first clock line, where the first capacitor is configured to receive a first clock signal having a first phase via the first clock line and couple a first bias current to the first JJ, and (2) a second JJ coupled via a second capacitor to a second clock line, where the second capacitor is configured to receive a second clock signal having a second phase via the second clock line and couple a second bias current to the second JJ. The superconducting IC further includes a first trap circuit for the first superconducting circuit and a second trap circuit for a second superconducting circuit having additional JJs.

Abstract: A method, computer program product, and computing system for generating a plurality of acoustic relative transfer functions associated with a plurality of audio acquisition devices of an audio recording system deployed in an acoustic environment. Acoustic relative transfer functions of at least a pair of audio acquisition devices of the plurality of audio acquisition devices may be compared. Location information associated with an acoustic source within the acoustic environment may be determined based upon, at least in part, the comparison of the acoustic relative transfer functions of the at least a pair of audio acquisition devices of the plurality of audio acquisition devices.

Abstract: Innovations in adaptive encoding and decoding for units of a video sequence can improve coding efficiency. For example, some of the innovations relate to encoding/decoding that includes adaptive switching of color spaces between units within a video sequence. Other innovations relate encoding/decoding that includes adaptive switching of color sampling rates between units within a video sequence. Still other innovations relate encoding/decoding that includes adaptive switching of bit depths between units within a video sequence.

Abstract: Methods, systems, and media are shown for creating and deploying honey tokens for intrusion detection in a remote computing resource system. Resource modules provisioned for a tenant are identified for intrusion detection. For each identified resource modules, a provisioned resource having a corresponding access credential is allocated and the access credential is deployed in the identified resource module. A data entry is created in a token mapping store that identifies the access credential and the resource module. Access logs are scanned to detect access attempts. For each access attempt, the token mapping store is searched for a data entry with an access credential that matches the access credential of the access attempt. If found, an alert is generated that includes the identified resource module of the matching data entry.

Abstract: A hinged mobile computing device includes a first housing part and a second housing part coupled by a hinge assembly having a harness, a harness cover, a first hinge body, and a second hinge body. The harness is configured to accommodate flexible printed circuitry and a cable that extend from the first housing part to the second housing part via the hinge assembly. The hinge bodies include respective friction bands, each friction band being configured to engage a respective shaft formed on the harness and having a gear configured to mesh with a respective cog arranged within the harness cover to coordinate a timing of the rotation of the first and second housing parts between face-to-face and back-to-back orientations. The hinge assembly further includes a spring-loaded opening mechanism.

Abstract: Disclosed in some examples are methods, systems, machine-readable mediums, and devices for providing an aggregated view of reactions to content across multiple sharing locations of a network-based content sharing service and in some examples across multiple sharing locations of multiple network-based content sharing services. The system may track or determine reactions across different sharing locations of a single content sharing service (such as a unified communication service) or across multiple services. In some examples, only reactions to the content that are from sharing locations accessible by a particular user are aggregated and displayed to the particular user. In other examples, reactions may be viewable from all sources regardless of access by a user.

Abstract: A computer-implemented method for performing natural language-based data integration includes causing execution of a data integration application on a remote device via a network and causing surfacing of a GUI corresponding to the data integration application on a display of the remote device. The method includes receiving, via the GUI, a natural language input representing a data integration task, generating, via an LLM, a set of ordered activities corresponding to the data integration task represented by the natural language input, and selecting, via the LLM, one or more APIs for performing each activity within the set of ordered activities. The method also includes generating a data pipeline based on the set of ordered activities and the API(s) for performing each activity, as well as back-translating the data pipeline to a desired data format for execution by the data integration application.

Abstract: In various embodiments, a computer-implemented method of training a neural network for creating an output signal of different modality from an input signal is described. In embodiments, the first modality may be a sound signal or a visual image and where the output signal would be a visual image or a sound signal, respectively. In embodiments a model is trained using a first pair of visual and audio networks to train a set of codebooks using known visual signals and the audio signals and using a second pair of visual and audio networks to further train the set of codebooks using the augmented visual signals and the augmented audio signals. Further, the first and the second visual networks are equally weighted and where the first and the second audio networks are equally weighted.

Abstract: In an example embodiment, an embedding model is used to generate an embedding of a natural language searching goal specified by a user, the embedding representing user intent of the user. Playbooks in a database of playbooks are also run through the embedding model to generate an embedding for each playbook indicative of a meaning of each playbook. A semantic relationship score can then be computed for each combination of the natural language search goal and a playbook, using the embeddings. These semantic relationship scores can then be passed into a ranking machine learning model, along with measured success rates for the playbooks, to generate a ranking of the playbooks. Based on this ranking, a set of filters and action corresponding to at least one of the playbooks may then be recommended to the user.

Abstract: Innovations in encoder-side options for intra block copy (“BC”) prediction mode facilitate intra BC prediction that is more effective in terms of rate-distortion performance and/or computational efficiency of encoding. For example, some of the innovations relate to concurrently performing block vector (“BV”) estimation and making block splitting decisions for a block. Other innovations relate to selectively merging blocks into a larger block during BV estimation.

Abstract: Systems and methods for spatial-textual clustering-based recognition of text in videos are disclosed. A method includes performing textual clustering on a first subset of a set of predictions that correspond to numeric characters only and performing spatial-textual clustering on a second subset of the set of predictions that correspond to alphabetical characters only. The method includes, for each cluster of predictions associated with the first subset of the set of predictions, choosing a first cluster representative to correct any errors in each cluster of predictions associated with the first subset of the set of predictions and outputting any recognized numeric characters. The method includes, for each cluster of predictions associated with the second subset of the set of predictions, choosing a second cluster representative to correct any errors in each cluster of predictions associated with the second subset of the set of predictions and outputting any recognized alphabetical characters.

Abstract: This document relates to providing meaningful information relating to a dataset. One example can obtain aggregated summaries and a related knowledge graph. The example can enable local, community, and global retrieval augmented generation utilizing the aggregated summaries and the knowledge graph.