Automatically recognizing and surfacing important moments in multi-party conversations

1. A non-transitory computer-readable medium comprising instructions encoded thereon to identify a moment in a transcript, the instructions when executed by at least one processor causing the at least one processor to:

receive a transcription of a conversation;

identify a participant of the conversation;

determine, based on text of the transcription, a stage corresponding to the transcription;

modify weights of a machine learning model corresponding to the participant based on the stage;

access the machine learning model corresponding to the participant, the machine learning model having the modified weights;

apply, as input to the machine learning model, the transcription;

receive, as output from the machine learning model, a portion of the transcription having relevance to the participant; and

generate for display, to the participant, information pertaining to the portion.

2. The non-transitory computer-readable medium of claim 1 , wherein the instructions to access the machine learning model corresponding to the participant further comprise instructions that when executed causes the at least one processor to:

determine a group of which the participant is a part;

identify a group model trained based on preferences of the group; and

assign the group model as the machine learning model corresponding to the participant.

3. The non-transitory computer-readable medium of claim 2 , wherein the group model is used to surface moments to other users who are a part of the group.

4. The non-transitory computer-readable medium of claim 1 , wherein each respective participant of the conversation has a separate respective machine learning model trained, using profile information of the respective participant for identification of one or more portions of transcriptions having relevance to that respective participant.

5. The non-transitory computer-readable medium of claim 1 , further comprising instructions to train the machine learning model, the instructions when executed causing the at least one processor to:

access a profile of the participant, the profile indicating terms in historical search queries performed by the participant, and indicating participant interaction with results of the historical search queries; and

label the terms based on the indicated participant interaction.

6. The non-transitory computer-readable medium of claim 5 , wherein a strength of association between the terms and the label is updated based on:

a frequency with which the participant uses the terms; and

how recently, relative to a present time, a term was used in the historical search queries by the participant.

7. The non-transitory computer-readable medium of claim 1 , wherein the instructions to apply, as input to the machine learning model, the transcription further comprise instructions that when executed causes the at least one processor to:

identify one or more word embeddings corresponding to the transcription; and

apply, as additional input to the machine learning model, the one or more word embeddings.

8. The non-transitory computer-readable medium of claim 1 , wherein the instructions to receive, as output from the machine learning model, a portion of the transcription having relevance to the participant further comprise instructions that when executed cause the at least one processor to:

receive a plurality of scores for different portions of the transcription;

compare each score of the plurality of scores to a threshold; and

determine the portion of the transcription having relevance to the participant based on its corresponding score exceeding the threshold.

9. The non-transitory computer-readable medium of claim 1 , further comprising instructions to receive the transcription automatically and in real-time during the conversation.

10. The non-transitory computer-readable medium of claim 9 , wherein the instructions to generate for display, to the participant, information pertaining to the portion further comprise instructions to display the information pertaining to the portion while the conversation continues to occur.

11. A method for identifying a moment in a transcript, the method comprising:

receiving a transcription of a conversation;

identifying a participant of the conversation;

determining, based on text of the transcription, a stage corresponding to the transcription;

modifying weights of a machine learning model corresponding to the participant based on the stage;

accessing the machine learning model corresponding to the participant, the machine learning model having the modified weights;

applying, as input to the machine learning model, the transcription;

receiving, as output from the machine learning model, a portion of the transcription having relevance to the participant; and

generating for display, to the participant, information pertaining to the portion.

12. The method of claim 11 , wherein accessing the machine learning model corresponding to the participant comprises:

determining a group of which the participant is a part;

identifying a group model trained based on preferences of the group; and

assigning the group model as the machine learning model corresponding to the participant.

13. The method of claim 11 , wherein each respective participant of the conversation has a separate respective machine learning model trained, using profile information of the respective participant, to identify one or more portions of transcriptions having relevance to that respective participant.

14. The method of claim 11 , wherein applying, as input to the machine learning model, the transcription comprises:

identifying one or more word embeddings corresponding to the transcription; and

applying, as additional input to the machine learning model, the one or more word embeddings.

15. The method of claim 11 , wherein receiving, as output from the machine learning model, a portion of the transcription having relevance to the participant comprises:

receiving a plurality of scores for different portions of the transcription;

comparing each score of the plurality of scores to a threshold; and

determining the portion of the transcription having relevance to the participant based on its corresponding score exceeding the threshold.

16. A system for identifying a moment in a transcript, the system comprising:

a first module for receiving a transcription of a conversation;

a second module for identifying a participant of the conversation;

a third module for determining, based on text of the transcription, a stage corresponding to the transcription, modifying weights of a machine learning model corresponding to the participant based on the stage, accessing the machine learning model corresponding to the participant, the machine learning model having the modified weights, applying, as input to the machine learning model, the transcription and receiving, as output from the machine learning model, a portion of the transcription having relevance to the participant; and

a fourth module for generating for display, to the participant, information pertaining to the portion.

17. The system of claim 16 , wherein each respective participant of the conversation has a separate respective machine learning model trained, using profile information of the respective participant, to identify one or more portions of transcriptions having relevance to that respective participant.

18. The system of claim 16 , wherein the machine learning model is trained by:

accessing a profile of the participant, the profile indicating terms in historical search queries performed by the participant, and indicating participant interaction with results of the historical search queries; and

labeling the terms based on the indicated participant interaction.