Natural language understanding system and dialogue systems

1. A natural language understanding (NLU) system for determining a meaning of an input text, comprising:

a first-level NLU sub-system, configured to receive the input text; and

at least one second-level NLU sub-system, each communicatively coupled with, and having a relatively higher performance than, the first-level NLU sub-system;

wherein:

the first-level NLU sub-system is configured, upon receiving the input text, to perform a first calculation over the input text, and then:

if the first calculation generates a first meaning corresponding to the input text with a first confidence level surpassing a first threshold, to output the first meaning; or

if otherwise, to pass on the input text to one of the at least one second-level NLU sub-system based on a pre-determined rule;

and

each of the at least one second-level NLU sub-system is configured, upon receiving the input text from the first-level NLU sub-system, to perform a second calculation over the input text, and then:

if the second calculation generates a second meaning corresponding to the input text with a second confidence level surpassing a second threshold, to output the second meaning; or

if otherwise, to output a result indicating a rejection of meaning.

2. The natural language understanding (NLU) system of claim 1 , wherein each of the at least one second-level NLU sub-system has a relatively richer user interface than the first-level NLU sub-system, wherein:

each of the at least one second-level NLU sub-system is further configured to periodically receive a first manual update for NLU elements; and

the first-level NLU sub-system is further configured to periodically receive a first automatic update for NLU elements from each of the at least one second-level NLU sub-system.

3. The natural language understanding (NLU) system of claim 2 , wherein:

the first-level NLU sub-system is further configured to receive a second manual update for NLU elements; and

each of the at least one second-level NLU sub-system is further configured to periodically receive a second automatic update for NLU elements from the first-level NLU sub-system.

4. The natural language understanding (NLU) system of claim 1 , wherein the performance is evaluated from an F1 standpoint.

5. The natural language understanding (NLU) system of claim 1 , wherein at least one of the first meaning and the second meaning is in a format of intent and slot.

6. The natural language understanding (NLU) system of claim 1 , wherein the first-level NLU sub-system comprises:

a first calculation module, configured to perform the first calculation over the input text to thereby generate a first result from the first calculation; and

a first determination module, configured to receive the first result from the first calculation module, and then to determine whether the first result comprises the first meaning.

7. The natural language understanding (NLU) system of claim 1 , wherein each of the at least one second-level NLU sub-system comprises:

a second calculation module, configured to perform the second calculation over the input text to thereby generate a second result from the second calculation; and

a second determination module, configured to receive the second result from the second calculation module, and then to determine whether the second result comprises the second meaning.

8. The natural language understanding (NLU) system of claim 2 , wherein each of the at least one second-level NLU sub-system comprises a first user interface, configured to allow a first human agent to manually update configuration of NLU elements in the each of the at least one second-level NLU sub-system therethrough.

9. The natural language understanding (NLU) system of claim 2 , wherein:

each of the at least one second-level NLU sub-system comprises a second user interface; and

the first-level NLU sub-system comprises a third user interface;

wherein:

the second user interface and the third user interface are together configured to allow a first automatic periodic update of configuration of NLU elements from the each of the at least one second-level NLU sub-system to the first-level NLU sub-system.

10. The natural language understanding (NLU) system of claim 9 , wherein:

the first-level NLU sub-system further comprises a fourth user interface, configured to allow a second human agent to manually update configuration of NLU elements in the first-level NLU sub-system therethrough; and

the second user interface and the third user interface are together further configured to allow a second automatic periodic update of configuration of NLU elements from the first-level NLU sub-system to the each of the at least one second-level NLU sub-system.

11. The natural language understanding (NLU) system of claim 1 , wherein a number of the at least one second-level NLU sub-system is one.

12. The natural language understanding (NLU) system of claim 1 , wherein a number of the at least one second-level NLU sub-system is more than one.

13. The natural language understanding (NLU) system of claim 12 , wherein each of the at least one second-level NLU sub-system is specific to a different domain, and the pre-determined rule is based on domain information associated with the input text.

14. The natural language understanding (NLU) system of claim 12 , wherein each of the at least one second-level NLU sub-system is specific to a different system, and the pre-determined rule is based on system information associated with the input text.

15. The natural language understanding (NLU) system of claim 12 , wherein the pre-determined rule is based on availability of each of the at least one second-level NLU sub-system.

16. A dialogue systems, comprising a natural language understanding (NLU) component, wherein the natural language understanding (NLU) component comprises a natural language understanding (NLU) system according to claim 1 .

17. The dialogue systems according to claim 16 , further comprising an input decoder, configured to convert a user input into the input text for the natural language understanding (NLU) component.

18. The dialogue systems according to claim 17 , wherein the input decoder comprises an automatic speech recognizer (ASR) component, configured to convert a speech from a user into the input text.

19. The dialogue systems according to claim 17 , wherein the input decoder comprises an optical character recognizer (OCR) component, configured to convert characters in an image into the input text.

20. The dialogue systems according to claim 16 , wherein the natural language understanding (NLU) component comprises:

an electronic device, provided with the first-level NLU sub-system of the natural language understanding (NLU) system; and

at least one server, respectively provided with the at least one second-level NLU sub-system of the natural language understanding (NLU) system;

wherein:

the electronic device is communicatively connected with the at least one server to thereby allow communication between the first-level NLU sub-system and the at least one second-level NLU sub-system.

21. The dialogue systems according to claim 20 , wherein the electronic device is a smartphone, a smart speaker, a tablet computer, a laptop, or a personal computer.

22. The dialogue systems according to claim 20 , wherein each of the at least one server is in a cloud.