Abstract: At a machine learning service, a data structure generated during the training phase of a machine learning model, as well as an input records associated with a result of the model, are analyzed. A first informational data set pertaining to the result, which indicates an alternative result, is generated. The first informational data set is transmitted to a presentation device with a directive to display a visual representation of the data set. In response to an exploration request pertaining to the first informational data set, a second informational data set indicating one or more observations of a training data set used for the model is transmitted to the presentation device.

Abstract: At an artificial intelligence system, a baseline set of informational content elements pertaining to an item for presentation to one or more potential item consumers is identified. One or more optimization iterations are implemented. In a particular iteration, a data set comprising interaction records of a target audience with the baseline set and with one or more variants of the baseline set is collected. Using the data set as input to a machine learning model, effectiveness metrics of the different informational elements are determined. A particular content element set to be presented to an audience is identified using the effectiveness metrics.

Abstract: Techniques for simulating a user in a conversation are described. A user simulation service and a conversation agent service conduct a dialog. The user simulation service compares a current sequence of stored labels corresponding to statements in the dialog with a plurality of candidate sequences of labels corresponding to statements in a plurality of candidate dialogs to identify a matching sequence of labels. The user simulation sequence identifies a base sequence of labels that includes the matching sequence of labels to identify a label corresponding to an act in the base sequence of labels following the matching sequence of labels. The user simulation service issues the act to the conversation agent service to be used as a simulated user act.

Abstract: Described are techniques for determining occurrence of an error within a system of related devices. If the device affected by the error is not known, other devices within the system may be used to determine whether a device is unresponsive, non-functional, or limited in functionality. The other devices may also be used to determine data indicative of characteristics of the error, such as statuses and activities of devices at or near the time of the error, which may be used to generate a query. The query may be used to determine possible control actions that may address the error or prevent subsequent errors.

Abstract: Technologies for determining, displaying, and leveraging reasons to buy for categories of items. Aspects (reasons to buy) may be predetermined for categories of items from a corpus of item information. For each item in a category, community-sourced content (e.g., customer reviews) may be identified for (aligned to) each aspect, and relevant comments may be extracted from the content. The comments may be analyzed, for example according to sentiment, to provide scoring and summary statistics for the aspects. Items may be ranked according to the summary statistics. The aspects for an item may be displayed, with extracted comments and/or summary statistics provided for each aspect. Items in a category or across categories may be compared according to the aspects. Aspects may be used in searches for items within categories to filter the items according to one or more of the aspects.