Selecting items from a database

A computer system for determining visual preference information for a user, the computer system comprising:

• one or more computer processors;
• a non-transitory computer readable medium storing a plurality of instructions, which when executed, cause the one or more processors to:
• select a display set of item images from the at least one item database;
• cause the display set of item images to be displayed at a user interface in an item array at respective array positions;
• detect engagement events at the user interface between the user and engaged-with item images of the display set;
• use the detected engagement events to determine a visual preference hypothesis for the user based on visual features extracted from the engaged-with item images;
• and
• update the display set with new item images selected from the at least one item database by comparing visual features extracted from the new item images with the visual preference hypothesis wherein the visual preference hypothesis corresponds to a type of visual feature or visual feature combination, and biases the selection of the new item images toward item images having visual features or combinations of visual features of that corresponding type which match the visual preference hypothesis but which exhibit variation in at least one additional visual feature type for testing the visual preference hypothesis, wherein the visual preference hypothesis is tested based on a level of user exposure to item images for testing the visual preference hypothesis and a level of user engagement with the item images for testing the visual preference hypothesis, and wherein the level of user exposure and the level of user engagement are stored as associated information with the visual preference hypothesis, the associated information embodied in a confidence measure associated with the visual preference hypothesis;
• wherein the display set is updated in dependence on the confidence measure, and the confidence measure is, in turn, updated as the user continues to engage with the new item images of the updated display set;
• wherein subsequent engagement events between the user and engaged-with new item images of the display set continue to be detected and used to iteratively refine the visual preference hypothesis as the user continues to interact with the display set, and the display set continues to be iteratively updated as the user continues to interact with the updated display set and the visual preference hypothesis is refined;
• wherein the engagement events comprise at least one active engagement event in which the user selects an item image of the display set at the user interface, wherein, in response to the active engagement event, at least one item image is removed from the display set, and replaced with one of the new item images displayed at the array position of the removed item image, and/or
• wherein the item array is zoomable and/or scrollable, and at least some of the new item images are selected dynamically in response to zoom events and/or scroll events at the user interface so as to dynamically populate previously unpopulated array positions within the item array based on the visual preference hypothesis.

A computer program product, comprising a non-transitory computer-readable medium having computer-readable program code embodied therein to be executed by one or more processors, the program code including instructions to:

• select a display set of item images from at least one item database;
• cause the display set of item images to be displayed at a user interface in an item array at respective array positions;
• detect engagement events at the user interface between a user and engaged-with item images of the display set;
• use the detected engagement events to determine a visual preference hypothesis for the user based on visual features of the engaged-with item images wherein the visual preference hypothesis corresponds to a type of visual feature or visual feature combination, and biases the selection of the new item images toward item images having visual features or combinations of visual features of that corresponding type which match the visual preference hypothesis but which exhibit variation in at least one additional visual feature type for testing the visual preference hypothesis, wherein the visual preference hypothesis is tested based on a level of user exposure to item images for testing the visual preference hypothesis and a level of user engagement with the item images for testing the visual preference hypothesis, and wherein the level of user exposure and the level of user engagement are stored as associated information with the visual preference hypothesis, the associated information embodied in a confidence measure associated with the visual preference hypothesis;
• wherein the display set is updated in dependence on the confidence measure, and the confidence measure is, in turn, updated as the user continues to engage with the new item images of the updated display set;
• and
• update the display set with new item images selected from the at least one item database by comparing visual features of the new item images with the visual preference hypothesis;
• wherein subsequent engagement events between the user and engaged-with new item images of the display set continue to be detected and used to iteratively refine the visual preference hypothesis as the user continues to interact with the updated display set, and the display set continues to be iteratively updated as the user continues to interact with the updated display set and the visual preference hypothesis is refined;
• wherein the engagement events comprise at least one active engagement event in which the user selects an item image of the display set at the user interface, wherein, in response to the active engagement event, at least one item image is removed from the display set, and replaced with one of the new item images displayed at the array position of the removed item image, and/or
• wherein the item array is zoomable and/or scrollable, and at least some of the new item images are selected dynamically in response to zoom events and/or scroll events at the user interface so as to dynamically populate previously unpopulated array positions within the item array based on the visual preference hypothesis.