Summarization of email on a client computing device based on content contribution to an email thread using classification and word frequency considerations

• US 10,353,994 B2
• Filed: 10/28/2016
• Issued: 07/16/2019
• Est. Priority Date: 11/03/2015
• Status: Active Grant

• Current Assignee: CommVault Systems Incorporated
• Original Assignee: CommVault Systems Incorporated
• Inventors: Upadhyay, Mrityunjay, Radhakrishnan, Divakar, Natarajan, Madhusudhanan
• Primary Examiner(s): Huynh, Cong-Lac
• Application Number:
  US15/337,897
• Publication Number:
  US 20170124034A1
• Time in Patent Office: 991 Days
• Field of Search: 715200, 715254, 715752
• CPC Class Codes: G06F 16/345 Summarisation for human users
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]
• CPC Class Codes: H04L 51/046 Interoperability with other network applications or services
• CPC Class Codes: H04L 51/216 Handling conversation history, e.g. grouping of messages in sessions or threads
• CPC Class Codes: H04L 51/42 Mailbox-related aspects, e.g. synchronisation of mailboxes

Abstract

Systems, methods, and computer-readable media are disclosed for enhancing an email application to automatically analyze an email thread and generate a compact content summary. The content summary is based on relative content contributions provided by the constituent email messages in the email thread. The content summary may be presented in a special window without disturbing or modifying the email thread or its constituent email messages. The distinctive content summary disclosed herein comprises certain sentences that are automatically gleaned from the email thread, analyzed relative to other sentences, and presented in a chronological sequence so that the user can quickly determine what the email thread is about and/or the current status of the conversation. The content summary is based on email weights, word weights, and intersecting sentence pairs.

Claims

1. A method comprising:

classifying, by a computing device executing an email application, a baseline set of email messages received and sent during a predefined baseline period of time, wherein each email message in the baseline set is classified as either important or not-important;

wherein a given email message in the baseline set is more likely to be classified as important rather than not-important, when the given email message was addressed personally to a user of the computing device executing the email application, and not to a group comprising the user;

by the computing device executing the email application, generating a first word-frequency table comprising words in email messages classified as important within the baseline set, wherein each word in the first word-frequency table is associated with a measure of frequency of occurrence of the word in the email messages classified as important;

by the computing device executing the email application, generating a second word-frequency table comprising words in email messages classified as not-important within the baseline set, which are not in the first word-frequency table, wherein each word in the second word-frequency table is associated with a measure of frequency of occurrence of the word in the email messages classified as not-important;

by the computing device executing the email application, generating a content summary of an email-thread comprising a plurality of email messages which are distinct from and created after the baseline set of email messages, wherein the content summary of the email-thread is generated based on, for each email message in the email-thread;

(i) an email-weight assigned to the email message in the email-thread, based at least in part on a proximity of the email message in the email-thread to a chronological median of the email-thread, and

(ii) a word-weight assigned to some words in the email message in the email-thread, wherein for any given word the word-weight is based at least in part on;

(a) a measure of frequency of the given word obtained from one of the first word-frequency table and the second word-frequency table,

(b) whether the given word is in the first word-frequency table versus in the second word-frequency table, and

(c) the email-weight of the email message in the email-thread comprising the given word.

2. The method of claim 1 further comprising:

by the computing device executing the email application, identifying a first set of intersecting sentence pairs from sentences in the email-thread,

wherein a first sentence in a given intersecting sentence pair has a second set of words in common with a second sentence in the given intersecting sentence pair; and

by the computing device executing the email application, assigning an intersection-score to each intersecting sentence pair in the first set, wherein the intersection-score is based on;

(a) word-weights of the words in the second set belonging to the first sentence in the given intersecting sentence pair, which are weighted by the email-weight of the email message in the email-thread comprising the first sentence, and

(b) word-weights of the words in the second set belonging to the second sentence in the given intersecting sentence pair, which are weighted by the email-weight of the email message in the email-thread comprising the second sentence.

3. The method of claim 2 further comprising:

by the computing device executing the email application, generating the content summary of the email-thread, which lists in chronological order unique sentences from highest-scoring intersecting sentence pairs.

4. The method of claim 1 wherein the given email message in the baseline set is further more likely to be classified as important rather than not-important when the given email message in the baseline set was read by the user of the computing device executing the email application.

5. The method of claim 1 wherein the given email message in the baseline set is further more likely to be classified as important rather than not-important when the user of the computing device executing the email application responded to the given email message in the baseline set.

6. The method of claim 1 wherein the given email message in the baseline set is further more likely to be classified as important rather than not-important when a follow-up flag was entered for the given email message in the baseline set.

7. The method of claim 1 wherein the word-weight for a given word is higher when the given word appears in the first word-frequency table, as compared to when the given word appears in the second word-frequency table.

8. The method of claim 1 further comprising:

by the computing device executing the email application, identifying a first set of intersecting sentence pairs among sentences in the email-thread,

wherein a first sentence in a given intersecting sentence pair has a second set of words in common with a second sentence in the given intersecting sentence pair; and

by the computing device executing the email application, generating the content summary of the email-thread, which lists in chronological order unique sentences from highest-scoring intersecting sentence pairs.

9. The method of claim 1 further comprising:

by the computing device executing the email application, identifying a first set of intersecting sentence pairs among sentences in the email-thread,

wherein a first sentence in a given intersecting sentence pair has a second set of words in common with a second sentence in the given intersecting sentence pair; and

by the computing device executing the email application, assigning an intersection-score to each intersecting sentence pair in the first set, wherein the intersection-score is based on;

(a) word-weights of the words in the second set belonging to the first sentence in the given intersecting sentence pair, and

(b) word-weights of the words in the second set belonging to the second sentence in the given intersecting sentence pair.

10. The method of claim 1 further comprising:

by the computing device executing the email application, generating the content summary of the email-thread, which lists in chronological order unique sentences from intersecting sentence pairs among sentences in the email-thread, wherein a first sentence in a given intersecting sentence pair has words in common with a second sentence in the given intersecting sentence pair.

11. A non-transitory computer-readable medium storing instructions that, when executed by a computing device comprising one or more processors and computer-readable memory, cause the computing device to perform operations comprising:

executing an email application;

classifying a baseline set of email messages received and sent during a predefined baseline period of time, wherein each email message in the baseline set is classified as either important or not-important;

wherein a given email message in the baseline set is more likely to be classified as important rather than not-important when the given email message was addressed personally to a user of the computing device executing the email application, and not to a group comprising the user;

generating a first word-frequency table comprising words in email messages classified as important within the baseline set, wherein each word in the first word-frequency table is associated with a measure of frequency of occurrence of the word in the email messages classified as important;

generating a second word-frequency table comprising words in email messages classified as not-important within the baseline set, which are not in the first word-frequency table, wherein each word in the second word-frequency table is associated with a measure of frequency of occurrence of the word in the email messages classified as not-important;

generating a content summary of an email-thread comprising a plurality of email messages which are distinct from and created after the baseline set of email messages, wherein the content summary of the email-thread is generated based on, for each email message in the email-thread;

(i) an email-weight assigned to the email message in the email-thread, based at least in part on a proximity of the email message in the email-thread to a chronological median of the email-thread, and

(ii) a word-weight assigned to some words in the email message in the email-thread, wherein for any given word the word-weight is based at least in part on;

(a) a measure of frequency of the given word obtained from one of the first word-frequency table and the second word-frequency table, and

(b) whether the given word is in the first word-frequency table versus in the second word-frequency table, and

(c) the email-weight of the email message in the email-thread comprising the given word.

12. The non-transitory computer-readable medium of claim 11, wherein the operations further comprise:

identifying a first set of intersecting sentence pairs from sentences in the email-thread,

wherein a first sentence in a given intersecting sentence pair has a second set of words in common with a second sentence in the given intersecting sentence pair; and

assigning an intersection-score to each intersecting sentence pair in the first set, wherein the intersection-score is based on;

(a) word-weights of words in the second set belonging to the first sentence in the given intersecting sentence pair, which are weighted by the email-weight of the email message in the email-thread comprising the first sentence, and

(b) word-weights of words in the second set belonging to the second sentence in the given intersecting sentence pair, which are weighted by the email-weight of the email message in the email-thread comprising the second sentence.

13. The non-transitory computer-readable medium of claim 12, wherein the operations further comprise:

generating the content summary of the email-thread, which lists in chronological order unique sentences from highest-scoring intersecting sentence pairs.

14. The non-transitory computer-readable medium of claim 11 wherein the given email message in the baseline set is further more likely to be classified as important rather than not-important when one or more of:

(i) the given email message in the baseline set was read by the user of the computing device executing the email application,

(ii) the user responded to the given email message in the baseline set, and

(iii) a follow-up flag was entered for the given email message in the baseline set.

15. The non-transitory computer-readable medium of claim 11 wherein the word-weight for a given word is higher when the given word appears in the first word-frequency table, as compared to when the given word appears in the second word-frequency table.

16. The non-transitory computer-readable medium of claim 11 wherein the operations further comprise:

generating the content summary of the email-thread, which lists in chronological order unique sentences from intersecting sentence pairs among sentences in the email-thread, wherein a first sentence in a given intersecting sentence pair has words in common with a second sentence in the given intersecting sentence pair.

17. A computing device comprising one or more processors and computer-readable memory, wherein the computing device is configured to:

execute an email application;

classify a baseline set of email messages received and sent during a predefined baseline period of time, wherein each email message in the baseline set is classified as either important or not-important;

wherein a given email message in the baseline set is more likely to be classified as important rather than not-important when the given email message in the baseline set was addressed personally to a user of the computing device and not to a group comprising the user;

generate a first word-frequency table comprising words in email messages classified as important within the baseline set, wherein each word in the first word-frequency table is associated with a measure of frequency of occurrence of the word in the email messages classified as important;

generate a second word-frequency table comprising words in email messages classified as not-important within the baseline set, which are not in the first word-frequency table, wherein each word in the second word-frequency table is associated with a measure of frequency of occurrence of the word in the email messages classified as not-important;

generate a content summary of an email-thread comprising a plurality of email messages which are distinct from and created after the baseline set of email messages, wherein the content summary of the email-thread is generated based on, for each email message in the email-thread;

(i) an email-weight assigned to the email message in the email-thread, based at least in part on a proximity of the email message in the email-thread to a chronological median of the email-thread, and

(ii) a word-weight assigned to some words in the email message in the email-thread, wherein for any given word the word-weight is based on one or more measures of frequency of the given word in one or more email messages in the email-thread.

18. The computing device of claim 17 wherein for any given word the word-weight is based at least in part on:

(a) a measure of frequency of the given word obtained from one of the first word-frequency table and the second word-frequency table,

(b) whether the given word is in the first word-frequency table versus in the second word-frequency table, and

(c) the email-weight of the email message in the email-thread comprising the given word.

19. The computing device of claim 17 wherein for any given word the word-weight is higher when the given word appears in the first word-frequency table, as compared to when the given word appears in the second word-frequency table.

20. The computing device of claim 17 wherein the computing device is further configured to:

generate the content summary of the email-thread, which lists in chronological order unique sentences from intersecting sentence pairs among sentences in the plurality of email messages in the email-thread, wherein a first sentence in a given intersecting sentence pair has words in common with a second sentence in the given intersecting sentence pair.