# US 2023/0056148 A1 · Personalized recommendation of complementary items to a user for inclusion in an order for fulfillment by an online concierge system based on embeddings for a user and for items

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0056148 A1
- Document ID: US20230056148A1
- Status: Published application
- Kind: A1
- Application number: 17/406,027
- Filed: 2021-08-18
- Published: 2023-02-23
- Earliest claimed: 2021-08-18
- Assignees: Maplebear, Inc.(dba Instacart)
- Applicants: Maplebear Inc.(dba Instacart)
- Inventors: Negin Entezari; Sharath Rao Karikurve; Shishir Kumar Prasad; Haixun Wang
- Web page: https://patentgrep.com/patent/US20230056148A1

## CPC

- G06Q 30/0633: {Managing shopping lists, e.g. compiling or processing purchase lists (shipping orders G06Q10/083; order filling G06Q10/087)}
- G06Q 10/087: Inventory or stock management, e.g. order filling, procurement or balancing against orders
- G06Q 10/0877: {by inventory control or reporting using inventory tracking or counting}
- G06Q 10/08744: {by distributed inventory management}
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

An online concierge shopping system identifies candidate items to a user for inclusion in an order based on prior user inclusion of items in orders and items currently included in the order. From a multi-dimensional tensor generated from cooccurrences of items in orders from various users, the online concierge system generates item embeddings and user embeddings in a common latent space by decomposing the multi-dimensional tensor. From items included in an order, the online concierge system generates an order embedding from item embeddings of the items included in the order. Scores for candidate items are determined based on similarity of item embeddings for the candidate items to the order embedding. Candidate items are selected based on their scores, with the selected candidate items identified to the user.

## Claims

### Claim 1

A method comprising:
obtaining, at an online concierge system, an inventory of items offered by one or more warehouses, the inventory identifying a warehouse offering an item and attributes of the item;
retrieving prior orders received from users of the online concierge system, each order including one or more items;
generating a tensor having more than two dimensions from the prior orders, each element of the tensor corresponding to a user, an item, and an additional item and having a value specifying a number of times the user corresponding to the identifier has included both the item and the additional item in one or more orders that were fulfilled by the online concierge system;
generating an item embedding for each item by decomposing the tensor;
generating a user embedding for each user by decomposing the tensor;
receiving a request from a user for an order, the request identifying the warehouse;
receiving selections of one or more items included in the order from the user;
generating an order embedding for the order from item embeddings for the one or more items included in the order;
identifying a set of candidate items offered by the warehouse identified by the order;
generating a score for each candidate item of the set, the score for a candidate item of the set based on the order embedding for the order and an item embedding of the candidate item;
selecting one or more candidate items based on the generated scores; and
transmitting an interface to a client device of the user including information describing the selected one or more candidate items.

### Claim 2 (depends on 1)

The method of claim 1, wherein generating the order embedding for the order from item embeddings for the one or more items included in the order comprises:
generating the order embedding as an average of the item embeddings for the one or more items included in the order.

### Claim 3 (depends on 2)

The method of claim 2, wherein generating the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises:
determining a product of the order embedding and the user embedding for the user;
determining the score for the candidate item of the set as a measure of similarity between the determined product and the item embedding for the candidate item of the set.

### Claim 4 (depends on 3)

The method of claim 3, wherein the measure of similarity between the determined product and the item embedding for the candidate item of the set comprises a dot product between the determined product and the item embedding for the candidate item of the set.

### Claim 5 (depends on 1)

The method of claim 1, wherein generating the order embedding for the order from item embeddings for the one or more items included in the order comprises:
generating the order embedding as a product of an average of the item embeddings for the one or more items included in the order and the user embedding.

### Claim 6 (depends on 5)

The method of claim 5, wherein generating the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises:
determining the score for a candidate item of the set as a measure of similarity between the order embedding and the item embedding for the candidate item of the set.

### Claim 7 (depends on 6)

The method of claim 6, wherein the measure of similarity between the determined product and the item embedding for the candidate item of the set comprises a dot product between the order embedding and the item embedding for the candidate item of the set.

### Claim 8 (depends on 1)

The method of claim 1, wherein selecting one or more candidate items based on the generated scores comprises:
ranking the candidate items of the set based on the generated scores; and
selecting one or more candidate items of the set having at least a threshold position in the ranking.

### Claim 9 (depends on 1)

The method of claim 1, wherein selecting one or more candidate items based on the generated scores comprises:
selecting one or more candidate items of the set having at least a threshold score.

### Claim 10 (depends on 1)

The method of claim 1, wherein identifying the set of candidate items offered by the warehouse identified by the order comprises:
identifying items offered by the order having item embeddings within a threshold distance from the order embedding.

### Claim 11

A computer program product comprising a non-transitory computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to:
obtain, at an online concierge system, an inventory of items offered by one or more warehouses, the inventory identifying a warehouse offering an item and attributes of the item;
retrieve prior orders received from users of the online concierge system, each order including one or more items;
generate a tensor having more than two dimensions from the prior orders, each element of the tensor corresponding to a user, an item, and an additional item and having a value specifying a number of times the user corresponding to the identifier has included both the item and the additional item in one or more orders that were fulfilled by the online concierge system;
generate an item embedding for each item by decomposing the tensor;
generate a user embedding for each user by decomposing the tensor;
receive a request from a user for an order, the request identifying the warehouse;
receive selections of one or more items included in the order from the user;
generate an order embedding for the order from item embeddings for the one or more items included in the order;
identify a set of candidate items offered by the warehouse identified by the order;
generate a score for each candidate item of the set, the score for a candidate item of the set based on the order embedding for the order and an item embedding of the candidate item;
select one or more candidate items based on the generated scores; and
transmit an interface to a client device of the user including information describing the selected one or more candidate items.

### Claim 12 (depends on 11)

The computer program product of claim 11, wherein generate the order embedding for the order from item embeddings for the one or more items included in the order comprises:
generate the order embedding as an average of the item embeddings for the one or more items included in the order.

### Claim 13 (depends on 12)

The computer program product of claim 12, wherein generate the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises:
determine a product of the order embedding and the user embedding for the user;
determine the score for the candidate item of the set as a measure of similarity between the determined product and the item embedding for the candidate item of the set.

### Claim 14 (depends on 13)

The computer program product of claim 13, wherein the measure of similarity between the determined product and the item embedding for the candidate item of the set comprises a dot product between the determined product and the item embedding for the candidate item of the set.

### Claim 15 (depends on 11)

The computer program product of claim 11, wherein generate the order embedding for the order from item embeddings for the one or more items included in the order comprises:
generate the order embedding as a product of an average of the item embeddings for the one or more items included in the order and the user embedding.

### Claim 16 (depends on 15)

The computer program product of claim 15, wherein generate the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises:
determine the score for a candidate item of the set as a measure of similarity between the order embedding and the item embedding for the candidate item of the set.

### Claim 17 (depends on 16)

The computer program product of claim 16, wherein the measure of similarity between the determined product and the item embedding for the candidate item of the set comprises a dot product between the order embedding and the item embedding for the candidate item of the set.

### Claim 18 (depends on 11)

The computer program product of claim 11, wherein select one or more candidate items based on the generated scores comprises:
rank the candidate items of the set based on the generated scores; and
select one or more candidate items of the set having at least a threshold position in the ranking.

### Claim 19 (depends on 11)

The computer program product of claim 11, wherein select one or more candidate items based on the generated scores comprises:
select one or more candidate items of the set having at least a threshold score.

### Claim 20 (depends on 11)

The computer program product of claim 11, wherein identifying the set of candidate items offered by the warehouse identified by the order comprises:
identify items offered by the order having item embeddings within a threshold distance from the order embedding.
