Abstract: The present disclosure relates generally to systems and methods for executing a search query using a combination of single-vector searching and multi-vector comparing to more efficiently and accurately identify a set of documents that are responsive to the search query. This involves applying a vector generation model to a search query to generate both single-vector and multi-vector embeddings for the search query. A single-vector search is performed to determine a candidate set of documents that might be relevant to the search query. A more robust multi-vector comparison is then performed to generate a result set including a re-ranked set of documents and/or a further reduced set of documents that is responsive to the search query.

Abstract: The present invention discloses a method for accounting carbon emissions during treating domestic waste leachate, and the method comprises the following steps: classifying treatment processes by types of waste treatment plants and corresponding leachate treatment processes; constructing a set of discharge factors in unit mass for each of treatment processes, and the factors include direct discharge factors and indirect discharge factors; obtaining total amount of carbon emissions in a region, based on the set of discharge factors, total treatment amount of waste leachate in the region, and the treatment ratio of each of treatment processes.

Abstract: Aspects disclosed herein include an electrochemical cell comprising: an anode comprising: a first surface comprising aluminum metal or an aluminum alloy; a liquid metal on the first surface, the liquid metal being in liquid state during operation of the battery and the liquid metal having a different composition than that of the first surface; and aluminum-rich dendrites extending from the first surface and in contact with an electrolyte; a positive electrode; and the electrolyte between the positive electrode and the negative electrode, the electrolyte being capable of conducting ions.

Abstract: A system, for example, an online system uses a machine learning based language model, for example, a large language model (LLM) to process crowd-sourced information provided by users. The crowd-sourced information may include comments from users represented as unstructured text. The system further receives queries from users and answers the queries based on the crowd-sourced information collected by the system. The system generates a prompt for input to a machine-learned language model based on the query. The system provides the prompt to the machine-learned language model for execution and receives a response from the machine-learned language model. The response comprises the insight on the topic and evidence for the insight. The evidence identifies one or more comments used to obtain the insight.

Abstract: Disclosed is a method for constructing a machine learning model for recommending an airway opening posture angle. The method includes collecting independent variable data affecting a glottis exposure effect when opening a patient airway, conducting an airway opening experiment, taking a posture angle corresponding to an optimal glottis exposure image as an optimal posture angle, performing single-factor analysis on the independent variables, and screening out statistically significant independent variable data as candidate prediction variables. The method further includes preselecting machine learning models, training and verifying the models, evaluating using AUC values, and selecting a model with a highest AUC value as a prediction model for recommending the optimal airway opening posture angle.

Abstract: A system may generate a prompt based in part on a search query from a customer client device. The prompt instructs a machine learned model to provide item predictions. And the model was trained by: converting structured data describing items of an online catalog to annotated text data (unstructured data), generating training examples based in part on the annotated text data, and training the model using the training examples. The system may receive item predictions generated by the prompt being applied to the machine learned model, the item predictions may have corresponding item identifiers. The item predictions are processed to identify a recommended item from the item predictions. The processing includes determining item information for the recommended item using an item identifier associated with the recommended item. The item information is provided to the customer client device.

Abstract: The present disclosure provides compounds of Formula (I), and pharmaceutically acceptable salts, solvates, hydrates, polymorphs, co-crystals, tautomers, stereoisomers, isotopically labeled derivatives, and prodrugs thereof. The provided compounds may be kinase (e.g., HCK, BTK, LYN, BLK, FRK) inhibitors. Also provided are pharmaceutical compositions and kits including the provided compounds. Further provided are methods of using the provided compounds, pharmaceutical compositions, and kits (e.g., for treating diseases (e.g., proliferative diseases) in a subject in need thereof).

Abstract: Provided in the present invention are inhibitors for programmed cell necrosis, a preparation method therefor and a use thereof. Specifically, provided in the present invention are a compound represented by formula I and a composition comprising same. The described compound may be used to prepare a pharmaceutical composition for the prevention and/or treatment of diseases involving cell death and/or inflammation.

Abstract: A system, for example, an online system uses a machine learning based language model, for example, a large language model (LLM) to process crowd-sourced information provided by users. The crowd-sourced information may include comments from users represented as unstructured text. The system further receives queries from users and answers the queries based on the crowd-sourced information collected by the system. The system generates a prompt for input to a machine-learned language model based on the query. The system provides the prompt to the machine-learned language model for execution and receives a response from the machine-learned language model. The response comprises the insight on the topic and evidence for the insight. The evidence identifies one or more comments used to obtain the insight.

Abstract: An online system uses benchmarking tests to identify indexing algorithms for an embedding database. To perform these benchmarking tests, the online system receives a set of parameters for configuring an embedding database. For example, the parameters may include a performance parameter and a latency parameter. The online system generates algorithm scores for a set of candidate indexing algorithms based on the parameters. Specifically, the online system tests each of the candidate indexing algorithms by generating a testing database based on a subset of the entries for the full database and by performing benchmarking tests on the testing database. The online system uses these tests to compute performance metrics for each candidate indexing algorithm and uses those performance metrics to compute an algorithm score for each indexing algorithm. The online system uses the computed algorithm scores to select an indexing algorithm for the embedding database.

Abstract: A vacuum cleaning system comprising a vacuum cleaner having a vacuum motor, a dirt bin, a primary separation system, a membrane filter, and an air valve arrangement configured to control air flow through at least the membrane filter. The vacuum cleaner is configured to be operable in a surface cleaning mode of operation and a self-cleaning mode of operation. In the surface cleaning mode of operation, the vacuum cleaner is configured such that the vacuum motor draws dirty air from a dirty-air inlet through the dirt bin, the primary separation system and the membrane filter in a first airflow direction, and, in the self-cleaning mode of operation, the vacuum cleaner is configured to permit air to flow from the air valve arrangement through the membrane filter in a second airflow direction to clean dirt from the membrane filter. Usefully, therefore, airflow can be routed through the vacuum cleaner in a second or ‘reverse’ direction to clean the caked on dirt from the membrane filter.

Abstract: Provided herein are methods of treating diseases (e.g., proliferative disease (e.g., cancer (e.g., breast cancer, colon cancer, testicular cancer, CNS cancer, stomach cancer, lymphoma (e.g., B-cell lymphoma (e.g., lymphoplasmacytic lymphoma (e.g., IgM secreting lymphoplasmacytic lymphoma (i.e., Waldenstrom's Macroglobulinemia), non-IgM secreting lymphoplasmacytic lymphoma)), diffuse large B-cell lymphoma (e.g., activated B-cell-like (ABC)-DLBCL, germinal center B-cell-like (GBC)-DLBCL), follicular lymphoma, marginal zone B-cell lymphoma, small lymphocytic lymphoma, mantle cell lymphoma), and leukemia (e.g., chronic lymphocytic leukemia (CLL), acute lymphoblastic leukemia, myelogenous leukemia (e.g., chronic myelogenous leukemia, acute myelogenous leukemia))))) comprising administering to the subject in need thereof a therapeutically effective amount of Compound (I). Further provided are methods for treating disease resistant to treatment with BTK inhibitors (e.g., ibmtinib).

Abstract: A computer system finetunes a machine-learned language model to generate a personalized response to a user request. The system may generate a user representation for each of a plurality of users by applying a transformer model to a sequence of tokens representing a sequence of activities of the user. The system may train an evaluation model coupled to receive a user representation and a response to a user request and generate an estimated evaluation score indicating a level of personalization of the response to the user. The system may finetune a first machine-learned language model to generate a second machine-learned language model. The finetuned machine-learned language model is configured to provide personalized responses for customer services at an online concierge system.

Abstract: An online system uses benchmarking tests to identify indexing algorithms for an embedding database. To perform these benchmarking tests, the online system receives a set of parameters for configuring an embedding database. For example, the parameters may include a performance parameter and a latency parameter. The online system generates algorithm scores for a set of candidate indexing algorithms based on the parameters. Specifically, the online system tests each of the candidate indexing algorithms by generating a testing database based on a subset of the entries for the full database and by performing benchmarking tests on the testing database. The online system uses these tests to compute performance metrics for each candidate indexing algorithm and uses those performance metrics to compute an algorithm score for each indexing algorithm. The online system uses the computed algorithm scores to select an indexing algorithm for the embedding database.

Abstract: A computer system finetunes a machine-learned language model to generate a personalized response to a user request. The system may generate a user representation for each of a plurality of users by applying a transformer model to a sequence of tokens representing a sequence of activities of the user. The system may train an evaluation model coupled to receive a user representation and a response to a user request and generate an estimated evaluation score indicating a level of personalization of the response to the user. The system may finetune a first machine-learned language model to generate a second machine-learned language model. The finetuned machine-learned language model is configured to provide personalized responses for customer services at an online concierge system.

Abstract: An online system receives a description of a task and generates a prompt for a generative machine-learned model. The prompt includes the description of the task and a request for a sequence of actions associated with the task, and a list of candidate actions and attributes related to the list of candidate actions. The online system provides the prompt to a model serving system deployed with the machine-learning model for execution. The online system obtains a response from the model serving system to extract wherein at least one workflow including the sequence of actions once executed completes a portion of the task. The online system receives, as output from the machine-learned model, the sequence of actions associated with the description of the task. The sequence of actions is executed in order to complete at least a portion of the task.

Abstract: This disclosure provides systems, methods and apparatuses for a high priority route selection descriptor (RSD) selection timer. A protocol data unit (PDU) session may be established for an application associated with a user equipment (UE). If the PDU session is associated with a lower RSD priority than the highest RSD priority for the application, the UE may initiate an RSD selection timer. After expiration of the RSD selection timer, the UE may try to establish another PDU session for the application with the highest RSD priority for the application. If another PDU session with the highest RSD priority is accepted, the UE may release the initial PDU session with the lower RSD priority and establish a new PDU session with the highest RSD priority. Such implementations may increase data rates, quality of service, and reliability for the application, decrease latency of the application, or decrease power consumption of the UE.

Abstract: A system may generate a prompt based in part on a search query from a customer client device. The prompt instructs a machine learned model to provide item predictions. And the model was trained by: converting structured data describing items of an online catalog to annotated text data (unstructured data), generating training examples based in part on the annotated text data, and training the model using the training examples. The system may receive item predictions generated by the prompt being applied to the machine learned model, the item predictions may have corresponding item identifiers. The item predictions are processed to identify a recommended item from the item predictions. The processing includes determining item information for the recommended item using an item identifier associated with the recommended item. The item information is provided to the customer client device.

Abstract: A system may generate a prompt based in part on a search query from a customer client device. The prompt instructs a machine learned model to provide item predictions. And the model was trained by: converting structured data describing items of an online catalog to annotated text data (unstructured data), generating training examples based in part on the annotated text data, and training the model using the training examples. The system may receive item predictions generated by the prompt being applied to the machine learned model, the item predictions may have corresponding item identifiers. The item predictions are processed to identify a recommended item from the item predictions. The processing includes determining item information for the recommended item using an item identifier associated with the recommended item. The item information is provided to the customer client device.

Abstract: The present disclosure provides a data acquisition and control system of pulsed power supply for multi-load, and relates to the technical field of pulsed power. The system of the present disclosure is configured to control 18 sets of pulsed power supplies to provide 18 coils with stable and reliable excitation current featuring multiple output time sequences and flexible adjustability, such that a near-earth space plasma environment ground-based simulation device can simulate the magnetic field topology of the earth magnetosphere and realize a background magnetic field required by physical experiments. The system mainly achieves the following functions: control over charge and discharge of pulsed power supplies, selection of power supplies put in use, “delay trigger” setting, voltage monitoring, status monitoring of key components and output current display; and it is also necessary to realize failure alarm and failure handling for the purpose of ensuring the safety of equipment and personnel.