Method and apparatus for summarization of dialogs

1. A method comprising:

obtaining an agent text stream and a customer text stream;

segmenting the agent text stream and customer text stream into sentences;

labeling sentences associated with the segmented agent text stream and the segmented customer text stream;

extracting sentences from the agent text stream and the customer text stream based upon frequencies of appearance of words and terms of interest;

generating an agent summary paragraph based on the extracted sentences from the agent text stream;

generating a customer summary paragraph based on the extracted sentences from the customer text stream; and

identifying keywords associated with the agent summary paragraph and the customer summary paragraph.

2. The method of claim 1 , wherein obtaining an agent text steam and customer text stream comprises:

receiving a transcript of a dialog between the agent and the customer;

identifying a first text stream within the transcript associated with the agent; and

identifying a second text stream within the transcript associated with the customer.

3. The method of claim 2 , wherein segmenting the agent text stream and customer text stream into sentences comprises:

assigning punctuation to the first text stream and the second text stream using a deep neural network model, wherein the deep neural network model includes at least one of a convolutional neural network model, a bidirectional long term memory model, or a transformer model.

4. The method of claim 1 , wherein labeling sentences associated with the agent text stream and the customer text stream comprises:

assigning an index to each sentence in the segmented agent text stream and to each sentence in the segmented customer text stream.

5. The method of claim 1 , wherein generating the agent summary paragraph and the customer summary paragraph comprises:

applying a summarization model to the segmented agent text stream and the segmented customer text stream, wherein the summarization model includes at least one of extractive summarization model based on a natural language tool kit, or an abstractive summarization model based on a transformer model.

6. The method of claim 5 , wherein applying the extractive summarization model comprises:

determining the terms of interest belonging to distinct subject classes;

assigning frequency related weights to each word in a segmented text stream, wherein particular weights are assigned to emphasize words that are the terms of interest in the segmented text stream.

7. The method of claim 6 , further comprising:

splitting words associated with vocabularies into a plurality of subsets, wherein each subset from the plurality of subsets is associated with a distinct subject matter class;

assigning weights to words corresponding to a frequency of an appearance of the words in the segmented text stream, wherein the assigned weights include particular weights to emphasize words associated with each subset;

assigning weights to sentences based on the assigned weights to words;

selecting a predetermined number of sentences based on the weights assigned to sentences; and

listing the selected sentences in an order to generate the summary paragraph.

8. The method of claim 1 , wherein identifying keywords associated with each of the agent summary paragraph and the customer summary paragraph further comprise:

finding keywords relevant to topics of interest, wherein topics of interest include technical information regarding equipment or services associated with customer.

9. The method of claim 1 , further comprising:

generating an interface for interacting with at least one of the agent summary paragraph or the customer summary paragraph based on the identified key words, wherein the interacting comprises:

detecting a selection of a key word in at least one of the agent summary paragraph or the customer summary paragraph;

identifying an index of a sentence in at least one of the segmented agent text stream or the segmented customer text stream based on the detected selection of the key word; and

displaying sentences corresponding to a predetermined range of indices from at least one of the segmented agent text stream or the segmented customer text stream, wherein the predetermined range of indices bound the identified index.

10. The method of claim 9 , wherein the interacting with at least one of the agent summary paragraph or the customer summary paragraph based on the identified key words comprises:

detecting a selection of a sentence in at least one of the agent summary paragraph or the customer summary paragraph;

identifying an index of the detected sentence in at least one of the segmented agent text stream or the segmented customer text stream based on the detected selection of the sentence; and

displaying neighboring sentences corresponding to a predetermined range of indices from at least one of the segmented agent text stream or the segmented customer text stream, wherein the predetermined range of indices bound the identified index associated with the selected sentence.

11. The method of claim 10 , wherein displaying neighboring sentences corresponding to a predetermined range of indices from at least one of the agent summary paragraph or the customer summary paragraph comprises:

determining a metric corresponding to lengths of the neighboring sentences; and

increasing a number of the neighboring sentences to display upon determining the metric corresponding to the length of the neighboring sentences is below a threshold.

12. The method of claim 9 , wherein the interacting with at least one of the agent summary paragraph or the customer summary paragraph based on the identified key words comprises:

detecting a selection of a key word associated with a first sentence in at least one of the agent summary paragraph or the customer summary paragraph;

identifying a predetermined number of sentences proximate to the first sentence in at least one of the segmented agent text stream or the segmented customer text stream, wherein the proximate sentences include the selected key word; and

displaying the first sentence and the predetermined number of sentences proximate to the first sentence.

13. A device comprising:

a communication interface;

a memory for storing instructions; and

a processor configured to execute the instructions to:

obtain an agent text stream and a customer text stream;

segment the agent text stream and customer text stream into sentences;

label sentences associated with the segmented agent text stream and the segmented customer text stream;

extract sentences from the agent text stream and the customer text stream based on frequencies of appearance of words and terms of interest;

generate an agent summary paragraph based on the extracted sentences from the agent text stream;

generate a customer summary paragraph based on the extracted sentences from the customer text stream; and

identify keywords associated with the agent summary paragraph and the customer summary paragraph.

14. The device of claim 13 , wherein the instructions to obtain an agent text steam and customer text stream further cause the processor to:

receive a transcript of a dialog between the agent and the customer;

identify a first text stream within the transcript associated with the agent; and

identify a second text stream within the transcript associated with the customer.

15. The device of claim 14 , wherein the instructions to segment the agent text stream and customer text stream into sentences further cause the processor to:

assign punctuation to the first text stream and the second text stream using a deep neural network model, wherein the deep neural network model includes at least one of a convolutional neural network model, a bidirectional long term memory model, or a transformer model.

16. The device of claim 13 , wherein the instructions to label sentences associated with the agent text stream and the customer text stream further cause the processor to:

assign an index to each sentence in the segmented agent text stream and to each sentence in the segmented customer text stream.

17. The device of claim 13 , wherein the instructions to generate the agent summary paragraph and the customer summary paragraph further cause the processor to:

apply a summarization model to the segmented agent text stream and the segmented customer text stream, wherein the summarization model includes at least one of extractive summarization based on a natural language tool kit, or an abstractive summarization based on a transformer model.

18. The device of claim 14 , further comprising instructions to:

generate an interface for interacting with at least one of the agent summary paragraph or the customer summary paragraph based on the identified key words, which further cause the processor to:

detect a selection of a key word in at least one of the agent summary paragraph or the customer summary paragraph;

identify an index of a sentence in at least one of the segmented agent text stream or the segmented customer text stream based on the detected selection of the key word; and

display sentences corresponding to a predetermined range of indices from at least one of the segmented agent text stream or the segmented customer text stream, wherein the predetermined range of indices bound the identified index.

19. The device of claim 18 , wherein the instructions to generate an interface for interacting with at least one of the agent summary paragraph or the customer summary paragraph based on the identified key words further cause the processor to:

detect a selection of a sentence in at least one of the agent summary paragraph or the customer summary paragraph;

identify an index of the detected sentence in at least one of the segmented agent text stream or the segmented customer text stream based on the detected selection of the sentence; and

display neighboring sentences corresponding to a predetermined range of indices from at least one of the segmented agent text stream or the segmented customer text stream, wherein the predetermined range of indices bound the identified index associated with the selected sentence.

20. A non-transitory computer-readable medium including instructions that, when executed by a processor of a device, cause the device to:

obtain an agent text stream and a customer text stream;

segment the agent text stream and customer text stream into sentences;

label sentences associated with the segmented agent text stream and the segmented customer text stream;

extract sentences from the agent text stream and the customer text stream based upon frequencies of appearance of words and terms of interest;

generate an agent summary paragraph based on the extracted sentences from the agent text stream;

generate a customer summary paragraph based on the extracted sentences from the customer text stream; and

identify keywords associated with each of the agent summary paragraph and the customer summary paragraph.