System and method for dynamically displaying images on electronic displays according to machine-learned patterns

A computer-implemented method for recommending at least one product to a user of a user device, comprising the steps of:

• receiving N product descriptions, each product description in the N product descriptions including an image;
• extracting, for each image in the N product descriptions, a plurality of features characterizing a visual quality of a product featured in the each image, the plurality of features including machine-understandable visual parameters, the visual parameters including shape of edges or visual patterns, the extracting including performing a computer vision analysis of the each image;
• arranging a subset of the N product descriptions relative to a surface in a grid;
• providing, to the user device, the subset of the N product descriptions;
• tracking, from the user device, signals indicating a sequence of interactions relative to the surface, each of the interactions including a micro-action among a plurality of micro-actions, the plurality of micro-actions including a scrolling action of the user relative to the grid and a click action relative to the grid;
• inferring, based on the tracking, one or more user preferences of the user of the user device and identifying one or more preferred features from among the plurality of extracted features that are associated with the one or more user preferences, the one or more user preferences comprising one or more product descriptions from among the subset of the N product descriptions;
• in response to the signals, rearranging the subset of the N product descriptions relative to the surface by:
• removing from an area in the grid arrangement one or more products proximate to a first product description in the subset of the N product descriptions,
• selecting, from the N product descriptions, replacement product descriptions to populate the area that contained the number of products removed,
• wherein at least one feature associated with the replacement product descriptions are related to features of the first product description,
• wherein at least one other feature associated with the replacement product descriptions represents a feature from among the one or more preferred features;
• selecting an escape item from a hypothesis testing repository and adding the new escape item to the subset of the N product descriptions, wherein the escape item represents a different preference that is not within the one or more preferred features, and wherein the escape item is represented as a new image;
• and
• monitoring how the user interacts or does not interact with the new image;
• wherein the signals indicating an interaction with the surface includes a click or touch on or selection of the first product description;
• the method further comprising the step of sending, to the user device, the rearranged subset of the N product descriptions including a product description including the new image.

A computer-implemented method for recommending at least one product to a user of a user device, comprising:

• receiving N product descriptions, each product description in the N product descriptions including an image configured to be selected by clicking, gesturing, or touching via a touch interface;
• extracting, for each image in the N product descriptions, a plurality of features including machine-understandable visual parameters, the extracting including performing a computer vision analysis of the each image to extract the machine-understandable visual parameters characterizing a visual quality of a product feature in the each image, the visual parameters including shape of edges or visual patterns;
• arranging a subset of the N product descriptions relative to a surface in a grid;
• providing, to the user device for display thereon, the subset of the N product descriptions;
• tracking, from the user device, signals indicating a sequence of interactions relative to the surface, each of the interactions including a micro-action among a plurality of micro-actions, the plurality of micro-actions including a scrolling action of the user relative to the grid and a click or touch action relative to the grid;
• inferring, based on the tracking, one or more user preferences of the user of the user device and identifying one or more preferred features from among the plurality of extracted features that are associated with the one or more user preferences, the one or more user preferences comprising one or more product descriptions from among the subset of the N product descriptions;
• in response to the signals, rearranging the subset of the N product descriptions relative to the surface, the rearranging including selecting a new image from a hypothesis testing repository and adding the new image to the subset of the N product descriptions, wherein the new image is an escape item representing a different preference that is not within the one or more user preferences;
• monitoring how the user interacts or does not interact with the escape item;
• and
• sending, to the user device, the rearranged subset of the N product descriptions including a product description including the new image.