Abstract: A computing system is described, where the computing system includes a processor and memory storing instructions that, when executed by the processor, cause the processor to perform several acts. The acts include receiving a query from an application executing on a client computing device that is in network communication with the computing system. The acts also include searching a computer-readable index of items based upon the query, identifying an item based upon the searching of the computer-readable index, transmitting the query to a computer-implemented model, and obtaining content generated by the computer-implemented model, where the computer-implemented model generated the content based upon the query. The acts further include returning at least one of the item or the content to the client computing device for presentment by way of the application executing on the client computing device.

Abstract: A computer-implemented technique performs machine learning that bypasses the traditional design of loss functions. The technique includes receiving plural instances of gradient objective information. Each of the plural instances includes a particular combination of plural gradient elements. The technique produces plural sets of machine-trained parameter values using the plural respective instances of gradient objective information. The technique performs this operation based on the plural instances of gradient objective information as given, without calculating the plural instances of gradient objective information using loss functions. The technique then measures performance of the plural sets of machine-trained parameter values in an application system. Based on the measured performance, the technique provides output information that identifies a particular set of machine-trained parameter values that satisfies a prescribed test.

Abstract: Methods and systems for providing targeted recommendations are provided. A user model is generated from user data. Feed candidates are generated based on the user model. The generated feed candidates are ranked on a predetermined scale. At least one targeted recommendation from the ranked feed candidates. The at least one targeted recommendation feed is provided to a user associated with the user data.

Abstract: Systems, methods, computer storage media, and user interfaces are provided for non-committal intent preview, disambiguation, and refinement of a search. A search prefix comprising one or more characters associated with an unexecuted search query is received. One or more intent suggestions are suggested to a user. For each of the one or more intent suggestions, one or more entity identifications associated with each of the one or more intent suggestions are received. Metadata corresponding to at least one entity associated with the one or more entity identifications is retrieved from an entity data store. Without retrieving search results for the unexecuted search query, an aggregated intent preview based on the retrieved metadata corresponding to the at least one entity is provided.

Abstract: Methods and systems for providing targeted recommendations are provided. A user model is generated from user data. Feed candidates are generated based on the user model. The generated feed candidates are ranked on a predetermined scale. At least one targeted recommendation from the ranked feed candidates. The at least one targeted recommendation feed is provided to a user associated with the user data.

Abstract: Systems, methods, computer storage media, and user interfaces are provided for non-committal intent preview, disambiguation, and refinement of a search. A search prefix comprising one or more characters associated with an unexecuted search query is received. One or more intent suggestions are suggested to a user. For each of the one or more intent suggestions, one or more entity identifications associated with each of the one or more intent suggestions are received. Metadata corresponding to at least one entity associated with the one or more entity identifications is retrieved from an entity data store. Without retrieving search results for the unexecuted search query, an aggregated intent preview based on the retrieved metadata corresponding to the at least one entity is provided.

Abstract: A method, system, and medium are provided for determining optimal sales rebate rates. Historical data, including sales data, price data, and rebate data are received, along with ongoing current data from current rebate transactions. Changes across the spectrum of data are determined and calculations are used to obtain an optimal sales rebate rate for one of more products or services utilizing statistical models, including but not limited to, a linear rebate rate model and a logarithmic-linear rebate rate model for one or more products or services. A mathematical analysis determines the appropriate model to use to obtain the optimal sales rebate rate. The optimal sales rebate rate may be applied to computing or non-computing environments, in whole or as a combination of both computing and non-computing environments.