Utilizing a dynamic memory network to track digital dialog states and generate responses

Patent metadata

Field	Value
Publication number	US-10909970-B2
Application number	US-201816135957-A
Country	US
Kind code	B2
Filing date	Sep 19, 2018
Priority date	Sep 19, 2018
Publication date	Feb 2, 2021
Grant date	Feb 2, 2021

Abstract

The present disclosure relates to generating digital responses based on digital dialog states generated by a neural network having a dynamic memory network architecture. For example, in one or more embodiments, the disclosed system provides a digital dialog having one or more segments to a dialog state tracking neural network having a dynamic memory network architecture that includes a set of multiple memory slots. In some embodiments, the dialog state tracking neural network further includes update gates and reset gates used in modifying the values stored in the memory slots. For instance, the disclosed system can utilize cross-slot interaction update/reset gates to accurately generate a digital dialog state for each of the segments of digital dialog. Subsequently, the system generates a digital response for each segment of digital dialog based on the digital dialog state.

First claim

What is claimed is:

1. A non-transitory computer readable storage medium comprising instructions that, when executed by at least one processor, cause a computing device to:
provide a segment of digital dialog to a dialog state tracking neural network comprising a dynamic memory network having a plurality of memory slots and a plurality of reset gates, wherein each memory slot of the plurality of memory slots corresponds to a designated dialog state characteristic;
utilize the dialog state tracking neural network to generate a digital dialog state corresponding to the segment of digital dialog by:
utilizing a reset gate associated with a first memory slot of the dynamic memory network to generate a value corresponding to a first designated dialog state characteristic for the first memory slot based on the segment of digital dialog, wherein the value replaces a previous value generated for the first memory slot based on a previous segment of digital dialog; and
generating the digital dialog state based on the value of the first memory slot that corresponds to the first designated dialog state characteristic; and

generate a digital response to the segment of digital dialog based on the digital dialog state.

2. The non-transitory computer readable storage medium of claim 1 , wherein the dynamic memory network further comprises a plurality of update gates corresponding to the plurality of reset gates and the plurality of memory slots.

3. The non-transitory computer readable storage medium of claim 2 , further comprising instructions that, when executed by the at least one processor, cause the computing device to further utilize the dialog state tracking neural network to generate the digital dialog state corresponding to the segment of digital dialog by utilizing an update gate associated with the first memory slot to determine an update value and apply the update value to modify an impact of the segment of digital dialog on the value.

4. The non-transitory computer readable storage medium of claim 2 , wherein utilizing the reset gate associated with the first memory slot to generate the value comprises determining a reset value and applying the reset value to modify an impact of the previous segment of digital dialog on the value.

5. The non-transitory computer readable storage medium of claim 1 , wherein:
the reset gate comprises a cross-slot interaction reset gate, and
utilizing the reset gate associated with the first memory slot to generate the value comprises:
comparing values of the plurality of memory slots to determine a cross-slot interaction reset value corresponding to the first memory slot; and
applying the cross-slot interaction reset value to modify an impact of the previous segment of digital dialog on the value.

6. The non-transitory computer readable storage medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to further utilize the dialog state tracking neural network to generate the digital dialog state corresponding to the segment of digital dialog by determining that the segment of digital dialog corresponds to a key vector associated with the first memory slot, and
wherein utilizing the reset gate associated with the first memory slot to generate the value comprises utilizing the reset gate to generate the value of the first memory slot further based on determining that the segment of digital dialog corresponds to the key vector associated with the first memory slot.

7. The non-transitory computer readable storage medium of claim 1 ,
further comprising instructions that, when executed by the at least one processor, cause the computing device to identify the segment of digital dialog by receiving an audio representation of the segment of digital dialog, and
wherein the digital response to the segment of digital dialog comprises an audio response.

8. The non-transitory computer readable storage medium of claim 1 ,
further comprising instructions that, when executed by the at least one processor, cause the computing device to further utilize the dialog state tracking neural network to generate the digital dialog state corresponding to the segment of digital dialog by:
generating a first dialog feature representation by processing the previous segment of digital dialog using one or more convolutional layers of the dialog state tracking neural network; and
generating a second dialog feature representation by processing the segment of digital dialog using the one or more convolutional layers of the dialog state tracking neural network, and
wherein generating the digital dialog state comprises generating the digital dialog state based on the first dialog feature representation, the second dialog feature representation, and the value of the first memory slot.

9. The non-transitory computer readable storage medium of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:
utilize the dialog state tracking neural network to generate the digital dialog state corresponding to the segment of digital dialog by utilizing the second dialog feature representation to generate a latent feature vector corresponding to the first memory slot; and
generate the value of the first memory slot further based on the latent feature vector.

10. The non-transitory computer readable storage medium of claim 1 , wherein the first designated dialog state characteristic comprises one of: a dialog topic; a location; an entity; or an action.

11. A system comprising:
at least one processor; and
at least one non-transitory computer readable storage medium storing instructions that, when executed by the at least one processor, cause the system to:
train a dialog state tracking neural network comprising a dynamic memory network having a set of memory slots and a set of cross-slot interaction reset gates corresponding to correlations between memory slots from the set of memory slots to generate digital dialog states used in generating digital responses to segments in digital dialogs by:
generating a first set of values for the set of memory slots based on a first training segment of training digital dialog;
applying a cross-slot interaction reset gate corresponding to a first memory slot from the set of memory slots based on cross-slot interactions between the first memory slot and other memory slots from the set of memory slots to generate a second set of values to replace the first set of values for the set of memory slots; and
generating a predicted dialog state for comparison with a ground truth dialog state based on the second set of values for the set of memory slots.

12. The system of claim 11 , wherein the dynamic memory network further comprises a set of cross-slot interaction update gates corresponding to the set of cross-slot interaction reset gates and the set of memory slots.

13. The system of claim 12 , further comprising instructions that, when executed by the at least one processor, cause the system to further train the dialog state tracking neural network to generate the digital dialog states by utilizing a cross-slot interaction update gate to determine a cross-slot interaction update value and apply the cross-slot interaction update value to modify an impact of a second training segment on the second set of values.

14. The system of claim 11 , further comprising instructions that, when executed by the at least one processor, cause the system to further train the dialog state tracking neural network to generate the digital dialog states by determining that a second training

Assignees

• Adobe Inc

Inventors

• Kim Seokhwan
• Chang Walter

Classifications

• G06N3/044 Primary

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/048

  Activation functions · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/09

  Supervised learning · CPC title