Granted Patent B2

US 12,450,517 · App. 18/155,443 · Granted Oct 21, 2025

Inventors: Atul Abhyankar (Pune, IN); Payal Panda (Bangalore, IN); Samanwita Majumdar (Kolkata, IN); Nishikanta Panda (Bangalore, IN); Swaroop Prasad Bhongade (Mumbai, IN); Shruti Sudhakar Marathe (Mumbai, IN); Shaista Firdose (Bangalore, IN)

Assignee: Accenture Global Solutions Limited

G06N10/60G06F3/0482G06F3/0484G06F40/134G06F40/166G06F40/20G06F40/40

Abstract

Claims (69)

1. A computer-implemented method for applying machine learning and quantum computing to generate carbon emissions for emails and modifying emails to reduce carbon emissions, the method comprising:

receiving an image of a draft email and an email log description of the draft email;

training a first classical machine learning model to classify draft emails including a single email address or a group email address in the “to:” field resulting in the first trained classical machine learning model;

applying the first trained classical machine learning model to the email log description to classify the draft emails as including the single email address or the group email address in the “to:” field;

training a second classical machine learning model to classify the draft emails as spam or standard emails resulting in the second trained classical machine learning model;

applying the second trained classical machine learning model to the email log description to classify the draft email as spam or a standard email;

training a QKNN machine learning model to classify an email as containing an attachment or not containing an attachment resulting in the trained QKNN machine learning model;

applying the trained quantum K nearest-neighbor (KNN) machine learning model to classify the draft email as containing the attachment or not containing the attachment;

based on output from the first trained classical machine learning model, the second trained classical machine learning model, and the trained QKNN machine learning model, determining one or more modifications that can be made to the draft email to reduce carbon emissions;

calculating a quantity of carbon emissions corresponding with each of the one or more modifications; and

presenting to a user, via a display of a user interface, the determined one or more modifications with the calculated quantity of carbon emissions.

2. The method of claim 1 , further comprising:

converting text of the email log description to text feature vectors;

inputting the text feature vectors into both first classical machine learning model and the trained second classical machine learning model, wherein applying the first trained classical machine learning model to the email log description includes applying the first trained classical machine learning model to the text feature vectors; and

applying the second trained classical machine learning model to the email log description includes applying the second trained classical machine learning model to the text feature vectors.

3. The method of claim 2 , further comprising:

prompting the user to select at least one of determined one or more modifications; and

upon selection of the at least one of determined one or more modifications, automatically implementing the modification to the draft email.

4. The method of claim 1 , wherein at least one of the determined one or more modifications is removal of one or more recipients in the draft email, further comprising:

modifying the draft email by removing one or more recipients in the draft email “to:” field; and

upon modification of the draft email, presenting to a user, via a display of a user interface, an amount of carbon emissions conserved by this removal and/or the quantity of carbon emissions corresponding to the modified draft email.

5. The method of claim 1 , wherein at least one of the determined one or more modifications is removal of one or more attachments in the draft email and replacing the removed attachments with a link to a shared location where files from the one or more attachments are saved.

6. The method of claim 5 , further comprising:

modifying the draft email by removing one or more attachments in a draft email and replacing the removed attachments with a link to a shared location where files from the one or more attachments are saved; and

upon modification of the draft email, presenting to a user, via a display of a user interface, an amount of carbon emissions conserved by this removal and replacement of the one or more attachments and/or the quantity of carbon emissions corresponding to the modified draft email.

7. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to:

receive an image of a draft email and an email log description of the draft email;

train a first classical machine learning model to classify draft emails including a single email address or a group email address in the “to:” field resulting in the first trained classical machine learning model;

apply the first trained classical machine learning model to the email log description to classify the draft emails as including the single email address or the group email address in the “to:” field;

train a second classical machine learning model to classify the draft emails as spam or standard emails resulting in the second trained classical machine learning model;

apply a second trained classical machine learning model to the email log description to classify the draft email as spam or a standard email;

train a QKNN machine learning model to classify an email as containing an attachment or not containing an attachment resulting in the trained QKNN machine learning model;

apply the trained quantum K nearest-neighbor (QKNN) machine learning model to classify the draft email as containing the attachment or not containing the attachment;

based on output from the first trained classical machine learning model, the second trained classical machine learning model, and the trained QKNN machine learning model, determine one or more modifications that can be made to the draft email to reduce carbon emissions;

calculate a quantity of carbon emissions corresponding with each of the one or more modifications; and

present to a user, via a display of a user interface, the determined one or more modifications with the calculated quantity of carbon emissions.

8. The non-transitory computer-readable medium storing software of claim 7 , wherein the instructions further cause the one or more computers to:

convert text of the email log description to text feature vectors;

input the text feature vectors into both first classical machine learning model and the trained second classical machine learning model, wherein applying the first trained classical machine learning model to the email log description includes applying the first trained classical machine learning model to the text feature vectors; and

apply the second trained classical machine learning model to the email log description includes applying the second trained classical machine learning model to the text feature vectors.

9. The non-transitory computer-readable medium storing software of claim 7 , wherein the instructions further cause the one or more computers to:

prompt the user to select at least one of determined one or more modifications; and

upon selection of the at least one of determined one or more modifications, automatically implement the modification to the draft email.

10. The non-transitory computer-readable medium storing software of claim 7 , wherein at least one of the determined one or more modifications is removal of one or more recipients in the draft email.

11. The non-transitory computer-readable medium storing software of claim 7 , wherein at least one of the determined one or more modifications is removal of one or more attachments in the draft email and replacing the removed attachments with a link to a shared location where the files from the one or more attachments are saved.

12. The non-transitory computer-readable medium storing software of claim 11 ,

wherein the instructions further cause the one or more computers to:

modify the draft email by removing one or more attachments in a draft email and replacing the removed attachments with a link to a shared location where the files from the one or more attachments are saved; and

upon modification of the draft email, present to a user, via a display of a user interface, an amount of carbon emissions conserved by this removal and replacement of the one or more attachments and/or the quantity of carbon emissions corresponding to the modified draft email.

13. A machine learning and quantum computing based system for reducing carbon emissions by modifying emails, the system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to:

receive an image of a draft email and an email log description of the draft email;

train a first classical machine learning model to classify draft emails including a single email address or a group email address in the “to:” field resulting in the first trained classical machine learning model;

apply the first trained classical machine learning model to the email log description to classify the draft emails as including the single email address or the group email address in a “to:” field;

train a second classical machine learning model to classify draft emails as spam or standard emails resulting in the second trained classical machine learning model;

apply the second trained classical machine learning model to the email log description to classify the draft email as spam or a standard email;

train a QKNN machine learning model to classify an email as containing an attachment or not containing an attachment resulting in the trained QKNN machine learning model

apply the trained quantum K nearest-neighbor (QKNN) machine learning model to classify the draft email as containing the attachment or not containing the attachment;

based on output from the first trained classical machine learning model, the second trained classical machine learning model, and the trained QKNN machine learning model, determine one or more modifications that can be made to the draft email to reduce carbon emissions;

calculate a quantity of carbon emissions corresponding with each of the one or more modifications; and

present to a user, via a display of a user interface, the determined one or more modifications with the calculated quantity of carbon emissions.

14. The system of claim 13 , wherein the instructions further cause the one or more computers to:

convert text of the email log description to text feature vectors;

input the text feature vectors into both first classical machine learning model and the trained second classical machine learning model, wherein applying the first trained classical machine learning model to the email log description includes applying the first trained classical machine learning model to the text feature vectors; and

apply the second trained classical machine learning model to the email log description includes applying the second trained classical machine learning model to the text feature vectors.

15. The system of claim 13 , wherein the instructions further cause the one or more computers to:

prompt the user to select at least one of determined one or more modifications; and

upon selection of the at least one of determined one or more modifications, automatically implement the modification to the draft email.

16. The system of claim 13 , wherein at least one of the determined one or more modifications is removal of one or more recipients in the draft email.

17. The system of claim 16 , wherein at least one of the determined one or more modifications is removal of one or more attachments in the draft email and replacing the removed attachments with a link to a shared location where the files from the one or more attachments are saved.

Assignments (2)

Continuity (1)

References Cited (12)