Increasing accuracy of delivery addresses

Patent metadata

Field	Value
Publication number	US-2026099805-A1
Application number	US-202519343413-A
Country	US
Kind code	A1
Filing date	Sep 29, 2025
Priority date	Oct 3, 2024
Publication date	Apr 9, 2026
Grant date	—

Abstract

Example implementations relate to improving an accuracy of an address. A set of address element inputs are received. A quality metric based on the set of address element inputs is computed using a first machine learning model. In accordance with a determination that the quality metric does not meet a threshold, an enriched address is generated based on the set of address element inputs. Feedback indicating an outcome of a delivery associated with the enriched address is received. The first machine learning model is re-train based on the feedback.

First claim

What is claimed is:

1 . A system, comprising:
a processor; and a non-transitory memory storing instructions, that when executed, cause the processor to:
receive a set of address element inputs via a first interface;
compute a quality metric based on the set of address element inputs using a first machine learning model;
in accordance with a determination that the quality metric does not meet a threshold, generate an enriched address based on the set of address element inputs;
receive feedback indicating an outcome of a delivery associated with the enriched address; and
re-train the first machine learning model based on the feedback.

2 . The system of claim 1 , wherein the set of address element inputs received via the first interface comprises a textual address and the instructions that cause the processor to generate the enriched address further comprise instructions to:
automatically convert the textual address into a location on a map; and display, on the map, a graphical user interface element indicative of the location.

3 . The system of claim 1 , wherein the instructions that cause the processor to generate the enriched address further comprise instructions to:
automatically determine latitudinal and longitudinal coordinates of a user providing the set of address element inputs; and display an interactive graphical user interface element that allows the user to obtain a user adjusted location for the enriched address based on movement of the interactive graphical user interface element.

4 . The system of claim 3 , wherein the quality metric comprises an address quality score computed at least in part based on a numerical value of a difference in distance between the user adjusted location and the determined latitudinal and longitudinal coordinates of the user providing the set of address element inputs.

5 . The system of claim 1 , wherein the first machine learning model used to compute the quality metric computed based on the set of address element inputs comprises a machine learning model trained on delivery failure outcomes comprising one or more of cancellations, delays or non-delivery.

6 . The system of claim 1 , wherein the instructions that cause the processor to generate the enriched address further comprise instructions to generate, using a recommendation model, a set of deliverable address choices that are generated based on geocoded, reverse-geocoded or past addresses associated with the set of address element inputs that resulted in successful deliveries.

7 . The system of claim 1 , wherein the first machine learning model is trained using a data source that comprises one or more of: geocoding inputs, input from third party databases, user provided delivery instructions, data from last mile deliveries, data signals indicating an outcome of a delivery, or changes made during deliveries.

8 . A computer-implemented method, comprising:
receiving a set of address element inputs via a first interface; computing a quality metric based on the set of address element inputs using a first machine learning model; in accordance with a determination that the quality metric does not meet a threshold, generating an enriched address based on the set of address element inputs; receiving feedback indicating an outcome of a delivery associated with the enriched address; and re-training the first machine learning model based on the feedback.

9 . The computer-implemented method of claim 8 , wherein the set of address element inputs received via the first interface comprises a textual address and generating the enriched address further comprises:
automatically converting the textual address into a location on a map; and displaying, on the map, a graphical user interface element indicative of the location.

10 . The computer-implemented method of claim 8 , wherein generating the enriched address further comprises:
automatically determining latitudinal and longitudinal coordinates of a user providing the set of address element inputs; and displaying an interactive graphical user interface element that allows the user to obtain a user adjusted location for the enriched address based on movement of the interactive graphical user interface element.

11 . The computer-implemented method of claim 10 , wherein the quality metric comprises an address quality score computed at least in part based on a numerical value of a difference in distance between the user adjusted location and the determined latitudinal and longitudinal coordinates of the user providing the set of address element inputs.

12 . The computer-implemented method of claim 8 , wherein the first machine learning model used to compute the quality metric computed based on the set of address element inputs comprises a machine learning model trained on delivery failure outcomes comprising one or more of cancellations, delays or non-delivery.

13 . The computer-implemented method of claim 8 , wherein generating the enriched address further comprises generating, using a recommendation model, a set of deliverable address choices that are generated based on geocoded, reverse-geocoded or past addresses associated with the set of address element inputs that resulted in successful deliveries.

14 . The computer-implemented method of claim 8 , wherein the first machine learning model is trained using a data source that comprises one or more of: geocoding inputs, input from third party databases, user provided delivery instructions, data from last mile deliveries, data signals indicating an outcome of a delivery, or changes made during deliveries.

15 . A non-transitory computer readable medium having instructions stored thereon, wherein the instructions, when executed by at least one processor, cause at least one device to perform operations comprising:
receiving a set of address element inputs via a first interface; computing a quality metric based on the set of address element inputs using a first machine learning model; in accordance with a determination that the quality metric does not meet a threshold, generating an enriched address based on the set of address element inputs; receiving feedback indicating an outcome of a delivery associated with the enriched address; and re-training the first machine learning model based on the feedback.

16 . The non-transitory computer readable medium of claim 15 , wherein the set of address element inputs received via the first interface comprises a textual address and the instructions that cause the at least one device to perform operations further comprise instructions to:
automatically convert the textual address into a location on a map; and display, on the map, a graphical user interface element indicative of the location.

17 . The non-transitory computer readable medium of claim 15 , wherein the instructions that cause the at least one device to perform operations further comprise instructions to:
automatically determine latitudinal and longitudinal coordinates of a user providing the set of address element inputs; and display an interactive graphical user interface element that allows the user to obtain a user adjusted location for the enriched address based on movement of the interactive graphical user interface element.

18 . The non-transitory computer readable medium of claim 17 , wherein the quality metric comprises an address quality score computed at least in part based on a numerical value of a difference in distance between the user adjusted location and the determined latitudinal and longitudina

Assignees

• Walmart Apollo Llc

Inventors

• Bendale Amit
• Sharan Anuja
• Suri Rakshit

Classifications

• G06Q10/0838

  Historical data · CPC title

• G06Q10/0843 Primary

  using forecasting or optimisation · CPC title