# US 2024/0256578 A1 · Systems and methods for cold-start recommendation using largescale graph models

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

- Number: US 2024/0256578 A1
- Document ID: US20240256578A1
- Status: Published application
- Kind: A1
- Application number: 18/104,184
- Filed: 2023-01-31
- Published: 2024-08-01
- Earliest claimed: 2023-01-31
- Assignees: Walmart Apollo, LLC
- Applicants: Walmart Apollo, LLC
- Inventors: Ankur Bhardwaj; Afroza Ali; Kannan Achan
- Web page: https://patentgrep.com/patent/US20240256578A1

## CPC

- G06F 16/288: {Entity relationship models}
- G06F 16/2237: {Vectors, bitmaps or matrices}
- G06F 16/9024: {Graphs; Linked lists (G06F16/9027 takes precedence)}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Systems and methods of generating interfaces including recommended items selected by a graph-based cold-start (GCS) model are disclosed. A request for an interface is received and a set of interface items is generated for inclusion in the interface. The set of interface items is selected, at least in part, by a GCS model including a semantic similarity component and a viewed-also-viewed component. The set of interface items is generated based on a combination of an output of the semantic similarity component and an output of the viewed-also-viewed component. The interface including the set of interface items is generated and transmitted to a system that generated the request for the interface.

## Claims

### Claim 1

A system, comprising:
a non-transitory memory;
a processor communicatively coupled to the non-transitory memory, wherein the processor is configured to read a set of instructions to:
receive a request for an interface;
generate a set of interface items for inclusion in the interface, wherein the set of interface items is selected at least in part by a graph-based cold-start (GCS) model including a semantic similarity component and a viewed-also-viewed component, and wherein the set of interface items is generated based on a combination of an output of the semantic similarity component and an output of the viewed-also-viewed component;
generate the interface including the set of interface items; and
transmit the interface to a system that generated the request for the interface.

### Claim 2 (depends on 1)

The system of claim 1, wherein the semantic similarity component is configured to generate a sparse distance matrix and the viewed-also-viewed component is configured to generate a co-viewed weighted sparse matrix.

### Claim 3 (depends on 2)

The system of claim 2, wherein the GCS model is configured to combine the sparse distance matrix and the co-viewed weighted sparse matrix to generate a product graph adjacency matrix.

### Claim 4 (depends on 2)

The system of claim 2, wherein the semantic similarity component is configured to generate a distance matrix, and wherein the sparse distance matrix is generated by modifying the distance matrix to remove values above a first predetermined threshold or below a second predetermined threshold.

### Claim 5 (depends on 2)

The system of claim 2, wherein the sparse distance matrix includes cosine distance values for each item pair in a catalog of items.

### Claim 6 (depends on 1)

The system of claim 1, wherein the GCS model is configured to generate an adjacency graph based on the combination of the output of the semantic similarity component and the output of the viewed-also-viewed component, and wherein the set of interface items is selected at least in part by a graph traversal process configured to traverse the adjacency graph.

### Claim 7 (depends on 6)

The system of claim 6, wherein the graph traversal process includes a Personalized PageRank process.

### Claim 8 (depends on 1)

The system of claim 1, wherein the GCS model is configured to generate a set of candidate items, wherein a ranking module is configured to generate a set of ranked items by ranking the set of candidate items, and wherein the set of interface items is selected in descending rank order from the set of ranked items.

### Claim 9 (depends on 1)

The system of claim 1, wherein the semantic similarity component is configured to generate a set of cold-start item recommendations and the viewed-also-viewed component is configured to generate a set of hot item recommendations.

### Claim 10

A computer-implemented method, comprising:
receiving a request for an interface;
generating, by a graph-based cold start (GCS) model, a sparse distance matrix and a co-viewed weighted matrix;
generating a set of interface items for inclusion in the interface, wherein the set of interface items is selected based at least in part on a combination of the sparse distance matrix and the co-viewed weighted matrix;
generating the interface including the set of interface items; and
transmitting the interface to a system that generated the request for the interface.

### Claim 11 (depends on 10)

The computer-implemented method of claim 10, wherein the sparse distance matrix is generated by a semantic similarity component of the GCS model and the co-viewed weighted sparse matrix is generated by a viewed-also-viewed component of the GCS model.

### Claim 12 (depends on 10)

The computer-implemented method of claim 10, comprising combining the sparse distance matrix and the co-viewed weighted sparse matrix to generate a product graph adjacency matrix, wherein the set of interface items is selected based at least in part on the graph adjacency matrix.

### Claim 13 (depends on 10)

The computer-implemented method of claim 10, wherein generating the sparse distance matrix comprises:
generating a distance matrix; and
removing values above a first predetermined threshold or below a second predetermined threshold from the distance matrix.

### Claim 14 (depends on 10)

The computer-implemented method of claim 10, wherein the sparse distance matrix includes cosine distance values for each item pair in a catalog of items.

### Claim 15 (depends on 10)

The computer-implemented method of claim 10, comprising generating, by the GCS model, an adjacency graph based on the combination of the combination of the sparse distance matrix and the co-viewed weighted matrix, and wherein the set of interface items is selected at least in part by a graph traversal process configured to traverse the adjacency graph.

### Claim 16 (depends on 15)

The computer-implemented method of claim 15, wherein the graph traversal process includes a Personalized PageRank process.

### Claim 17 (depends on 10)

The computer-implemented method of claim 10, wherein the GCS model is configured to generate a set of candidate items, wherein generating the set of interface items comprises generating, by a ranking model, a set of ranked items by ranking the candidate items, and wherein the set of interface items is selected in descending rank order from the set of ranked items.

### Claim 18 (depends on 10)

The computer-implemented method of claim 10, wherein the sparse distance matrix is representative of a set of cold-start item recommendations and the co-viewed weighted matrix is representative of a set of hot item recommendations.

### Claim 19

A computer-implemented method, comprising:
receiving a request for an interface;
generating, by a graph-based cold start model, a sparse distance matrix and a co-viewed weighted matrix;
combining the sparse distance matrix and the co-viewed weighted matrix to generate an adjacency matrix;
converting the adjacency matrix to a graph, wherein each node of the graph is representative of an item in a catalog of items and each edge in the graph is representative of a relationship value between the item and a second item in the catalog of items;
traversing the graph to identify a set of candidate items;
generating, by a ranking module, a set of ranked items by ranking the set of candidate items;
selecting a set of interface items for inclusion in the interface, wherein the set of interface items is selected from the set of ranked items in descending rank order;
generating the interface including the set of interface items; and
transmitting the interface to a system that generated the request for the interface.

### Claim 20 (depends on 19)

The computer-implemented method of claim 19, wherein the graph is traversed by a Personalization PageRank process.
