Augmenting an availability of an item at a warehouse determined from an availability model with information from shoppers fulfilling orders on item availability

A non-transitory computer readable storage medium for storing instructions, wherein the instructions, when executed by one or more processors, cause the one or more processors to:

• receive, at an online system, a request from a user to select items for fulfillment;
• determine, by the online system, availabilities of the items, wherein determining the availabilities of the items comprises applying a machine-learned availability model to predict the availabilities, wherein the machine-learned availability model is trained by inventory information and past outcomes of delivery orders;
• determine that an item has less than a threshold availability as predicted by the machine-learned availability model;
• transmit a request for a shopper to provide an indication of availability of the item;
• receive, at the online system from a client device of the shopper, a location-based image captured at a location corresponding to the item;
• automatically process the location-based image to detect the location of the item;
• and
• retrain the machine-learned availability model using the indication, wherein the retraining comprises incorporating updated availability information and the location-based image as additional training data such that subsequent availability predictions for the item are based on both historical inventory information and location-specific image data,
• wherein retraining the machine-learned availability model comprises:
• receiving the indication of the availability of the item from the client device of the shopper;
• and
• retraining the machine-learned item availability model in response to the indication indicating the item is available.

A system comprising:

• one or more processors;
• memory storing instructions, wherein the instructions, when executed by the one or more processors, cause the one or more processors to:
• receive, at an online system, a request from a user to select items for fulfillment;
• determine, by the online system, availabilities of the items, wherein determining the availabilities of the items comprises applying a machine-learned availability model to predict the availabilities, wherein the machine-learned availability model is trained by inventory information and past outcomes of delivery orders;
• determine that an item has less than a threshold availability as predicted by the machine-learned availability model;
• transmit a request for a shopper to provide an indication of availability of the item;
• receive, at the online system from a client device of the shopper, a location-based image captured at a location corresponding to the item;
• automatically process the location-based image to detect the location of the item;
• and
• retrain the machine-learned availability model using the indication, wherein the retraining comprises incorporating updated availability information and the location-based image as additional training data such that subsequent availability predictions for the item are based on both historical inventory information and location-specific image data,
• wherein retraining the machine-learned availability model comprises:
• receiving the indication of the availability of the item from the client device of the shopper;
• and
• retraining the machine-learned item availability model in response to the indication indicating the item is available.