Speech interactive system and method

• US 8,234,114 B2
• Filed: 08/14/2009
• Issued: 07/31/2012
• Est. Priority Date: 02/27/2009
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Chang, Yao-Yuan, Chang, Sen-Chia, Chien, Shih-Chieh, Tu, Jia-Jang
• Primary Examiner(s): Han, Qi
• Application Number:
  US12/541,872
• Publication Number:
  US 20100223060A1
• Time in Patent Office: 1,082 Days
• Field of Search: 704/251, 704/231, 704/235, 704/270, 704/257
• US Class Current: 704/252
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context

Abstract

The present invention relates to a speech interactive system and method. The system comprises a target information receiving module, an interactive mode setting and speech processing module, an interactive information update module, a decision module, and an output response module. It receives target information and sets corresponding target text sentence information. It also receives a user'"'"'s speech signal, sets an interactive mode, decides the speech'"'"'s target text sentence information, and generates an assessment for the target text sentence. Under the set interactive mode, the system updates the information in an interactive information recording table according to the assessment and a timing count. According to the interactive mode and the recorded information, an output mode for the target text sentence information is generated. According to the output mode and the recorded information, the response information is generated.

Claims

1. A speech interactive system implemented in a computer system having at least a processing device, a storage device, a text input device, a speech input device and an output device, comprising:

a target information receiving module implemented by said processing device and including said text input device for receiving target information and setting corresponding target text sentence information;

an interactive mode setting and speech processing module implemented by said processing device and including said speech input device for receiving a speech signal, setting an interactive mode, determining target text sentence information for said speech signal, and outputting an assessment for a target text sentence;

an interactive information update module implemented by said processing device for updating information in an interactive information recording table stored in said storage device according to said assessment for said target text sentence and a timing count;

a decision module implemented by said processing device for deciding an output response for said target text sentence information according to said interactive mode and said information in said interactive information recording table; and

an output response module implemented by said processing device and including said output device for outputting response information according to said output response and said information of said interactive information recording table.

2. The system as claimed in claim 1, wherein when only said speech signal is inputted to said interactive mode setting and speech processing module, said interactive mode is set as a conversation mode.

3. The system as claimed in claim 2, wherein said conversation mode is transmitted to said decision module.

4. The system as claimed in claim 2, wherein said interactive mode setting and speech processing module further includes a speech recognition unit for receiving said speech signal and determining target text sentence information of said speech signal.

5. The system as claimed in claim 1, wherein when said speech signal is inputted to said interactive mode setting and speech processing module and said target information is provided to said interactive mode setting and speech processing module, said interactive mode is set as a teaching mode.

6. The system as claimed in claim 5, wherein said teaching mode is transmitted to said interactive information update module.

7. The system as claimed in claim 6, wherein when said speech signal is determined as disqualified by an utterance verification method, said teaching mode is further transmitted to said decision module.

8. The system as claimed in claim 1, wherein said interactive information recording table records information of a speech interactive text sentence, said information of said speech interactive text sentence is chosen from any combination of an interactive text sentence, an accumulated teaching result, an assessment threshold, a difficulty level, and response information of said target text sentence.

9. The system as claimed in claim 8, wherein said system constructs a forgetting formula and said forgetting formula operates with a timing counter.

10. The system as claimed in claim 9, wherein a value of said forgetting formula is obtained via computation of said accumulated teaching result and said difficulty level.

11. The system as claimed in claim 1, wherein said response information at least includes correct response of said target text sentence, confused response of said target text sentence, and output response of a midway state with said target text sentence being started to be learned and the learning not completed yet.

12. The system as claimed in claim 1, wherein said system updates said interactive information recording table via an interactive information external update module.

13. The system as claimed in claim 12, wherein said system stores said interactive information recording table with a memory device, and updates said interactive information recording table by replacing said memory device.

14. The system as claimed in claim 12, wherein said interactive information external update module is installed on a computer system and linked to said speech interactive system to update said interactive information recording table of said speech interactive system.

15. The system as claimed in claim 1, wherein said system is embedded inside a body of a toy.

16. The system as claimed in claim 15, wherein said speech input device includes a microphone for receiving a speech input as said speech signal and said output device includes a speech player for outputting a speech response corresponding to said accumulated teaching result.

17. A speech interactive method implemented in a computer system having at least a processing device, a storage device, a text input device, a speech input device and an output device, comprising:

preparing an interactive information recording table in said storage device for recording speech interactive text sentence information;

receiving a speech signal from a user using said speech input device and determining target text sentence information for said speech signal, or receiving given target text sentence information as well using said text input device, in order to provide said user with different interactive modes for a speech interactive operation;

performing a speech assessment on said target text sentence information and generating an assessment result of a target text sentence;

updating said interactive information recording table according to said generated speech assessment result and a timing count if said target text sentence information is given and received from said text input device as well;

deciding an output response of said target text sentence information according to the interactive mode, speech assessment result, and said interactive information recording table; and

outputting response information using said output device according to said decided output response and said interactive information recording table.

18. The method as claimed in claim 17, wherein said different interactive modes includes a teaching mode and a conversation mode.

19. The method as claimed in claim 17, wherein said interactive information recording table records information of a speech interactive text sentence, and said information of said speech interactive text sentence is chosen from any combination of an interactive text sentence, an accumulated teaching result, an assessment threshold, a difficulty level, and response information of said target text sentence.

20. The method as claimed in claim 19, wherein said accumulated teaching result is related to a suspension time, said suspension time is a time unit passed since last update of learning said target text sentence, said accumulated teaching result is obtained via a forgetting formula, and a value of said forgetting formula is obtained via computation of said accumulated teaching result and said difficulty level.

21. The method as claimed in claim 17, wherein when only said speech signal is received from said speech input device, a conversation mode is provided to said user for an interactive operation, and all interactive text sentences in said interactive information recording table are used as recognition targets for recognizing target text sentence information corresponding to said speech signal, then a speech assessment is performed with the recognized target text sentence information and said speech signal to obtain a verification score.

22. The method as claimed in claim 21, said method refers to a combination of at least one previous said verification score, dynamically obtains a threshold adjustment value of speech verification and updates said interactive information recording table according to said threshold adjustment value.

23. The method as claimed in claim 17, wherein when said speech signal is received from said speech input device and said target text sentence information is given and received from said text input device as well, a teaching mode is provided to said user for said speech interactive operation, a speech assessment is performed on said target text sentence information to generate a speech assessment result, and said interactive information recording table is updated according to said speech assessment result.

24. The method as claimed in claim 17, wherein said response information at least includes correct response of said target text sentence, confused response of said target text sentence, and output response of a midway state with said target text sentence being started to be learned and the learning not completed yet.

25. The method as claimed in claim 24, wherein said output response of a midway state is to output a drawing or an image corresponding to said target text sentence as a response based on a ratio of accumulated teaching result/difficulty level, and for a speech output response, to output partial speech corresponding to said target text sentence according to said ratio of accumulated teaching result/difficulty level.