In various embodiments, a training application generates a preference prediction model based on an interaction matrix and a closed-form solution for minimizing a Lagrangian. The interaction matrix reflects interactions between users and items, and the Lagrangian is formed based on a constrained optimization problem associated with the interaction matrix. A service application generates a first application interface that is to be presented to the user. The service application computes predicted score(s) using the preference prediction model, where each predicted score predicts a preference of the user for a different item. The service application then determines a first item from the items to present to the user via an interface element included in the application interface. Subsequently, the service application causes a representation of the first item to be displayed via the interface element included in the application interface.

Search queries are received and search results are provided. Interaction tracking is used to determine with which search results users interact. The search results having received interactions can be represented as item vectors, which can include a vector representation of a portion of the search result, such as a title, description, or image. For each search query, the item vectors are aggregated, such as by averaging the item vectors. The search queries are stored in an item dataset as collected search queries respectively associated with the aggregate item vectors. When a new search query is received, a search query vector can be compared to the aggregate item description vectors to identify collected search queries that are related. The related collected search queries can be provided as search query recommendations or search results associated with the collected search queries can be provided in response to receiving the new search query.

A method for space design includes: receiving first information of a physical space by a terminal device, the first information including size measurement information of the physical space; acquiring second information of a target object, the second information including size measurement information of the target object; performing a visual presentation of the first information of the physical space and the second information of the target object on a display interface of the terminal device; calculating construction information of the target object being placed in the physical space by the terminal device based on a layout of the target object on the visual presentation according to the first information of the physical space and the second information of the target object; generating order information corresponding to the target object according to the calculated construction information.

Aspects of the present disclosure involve a system comprising a computer-readable storage medium storing a program and method for presenting cross-sell products. The program and method provide for receiving indication of a user selection to display cross sell data for at least one product made available for purchase by a website; determining a set of cross-sell products for the at least one product, each cross-sell product in the set of cross-sell products having been previously sold together with the at least one product in association with website; determining, for each cross-sell product in the set of cross-sell products, a set of metrics that relate the cross-sell product to the at least one product, and causing, for each cross-sell product in the set of cross-sell products, display of the respective set of metrics that relate the cross-sell product to the at least one product.

Described herein is a method and system for collection creation. The collection is a curated selection of product and/or services from an electronic catalogue presented to an online consumer. A repository of templates is provided. The first set of attributes of the items of catalog is determined by an attribute classification model. In parallel, data relevant to items from media sources are extracted. Features are identified from the extracted data. A second set of preferred attributes from the identified features is identified. A common set of attributes amongst the first and second set of attributes is identified. Templates that match a minimum number of the template's attributes with the item's attributes are selected. The selected templates are applied to the items of electronic catalog to automatically create curated collections. Curated collections are activated and displayed to the online consumer on a user interface.

Described herein is a method and system for creating templates. Themes are identified based on textual and image processing of media, wherein each of the themes is one or a combination of entities such as occasions, events, festivals, and seasonal wear. An attribute classification model is applied and thereafter attributes are identified based on textual and image processing of the media. Themes(s) are mapped to the attributes to create a template(s), wherein a map is a rule connecting themes to attributes. The template comprises a theme and its associated attributes, and the values of the attributes. A similarity search model is applied to create an extended range of templates from an initial template. The template is not a product and does not have a state. An extended range of templates are stored in a database.

Described herein is a system and computer implemented method of grouping clothing products by brands within a set of clothing images in an electronic catalog of an internet store serving online customers. Apply an object detection model to extract the dress section within the clothing image(s) to create preprocessed image(s). A machine learning model model is applied to the preprocessed image(s) to convert the image into a vector representation through an unsupervised technique. The vector contains the design features of the clothing image. The design features are representative of the brands. A clustering model is applied on the vector representations to arrive at the grouping of similar images of the clothing products. The grouped clothing products are displayed via a user interface, ordered by brands, to the online customers.

Methods and systems for governing execution of a geography dependent computer process are provided. In one aspect, a method includes receiving a target location comprising a geographic identifier. The method also includes accessing a data store comprising a plurality of maps. Each map of the plurality of maps includes a plurality of geometric shapes and each geometric shape is associated with an execution rule. The method also includes determining which geometric shape of the plurality of geometric shapes the geographic identifier is bounded within for at least two maps of the plurality of maps. The method further includes generating an aggregate rule set based on each execution rule associated with each geometric shape determined to bound the geographic identifier. The method further includes applying the aggregate rule set. Machine-readable media are also provided.

A system and method for providing proxy picking of non-fungible goods within an automated storage and retrieval system is provided, which repurposes one or more automated mobile robots operating within the automated inventory management system to perform a plurality of tasks across multiple different areas of an automated store. The proxy picking system and method are configured to pick individually identified non-fungible goods according to a customer selection on an ordering screen based on measured attributes and images of the goods, the attributes selected by the customer.

The description relates to an interactive physical product browsing experience. One example includes a display system for displaying objects. The display system comprises a plurality of object displays each comprising an object store, a robot for moving objects to and from a repository, a user input receiving means, and a processing means. The processing means is configured to cause the object displays to display some of the objects, monitor user behaviour using the user input receiving means, and cause the robot to move an object from the repository to one of the object stores based on the user behaviour.