Simultaneous translation of open domain lectures and speeches

1. A speech translation system for translating speech comprising:

a first-language automatic speech recognition unit configured for accepting spoken sound from a first speaker in a first language and for creating a plurality of partial hypotheses of the spoken sound of the first speaker in substantially real time while the first speaker is speaking;

a first-language resegmentation unit in communication with the first-language automatic speech recognition unit, wherein the first-language resegmentation unit is configured to:

merge at least two of the partial hypotheses received from the first-language automatic speech recognition unit; and

resegment the merged partial hypotheses into a first-language translatable segment in the first language, wherein a segment boundary for the first-language translatable segment is determined based on sound from a second speaker in a second language; and

a first-language machine translation unit, in communication with the first-language resegmentation unit, that receives the first-language translatable segment in the first language from the first-language resegmentation unit, wherein the first-language machine translation unit outputs a translation of the spoken sound from the first speaker into the second language based on the received first-language translatable segment.

2. The speech translation system of claim 1 , further comprising:

a second-language automatic speech recognition unit configured for accepting spoken sound from the second speaker in the second language and for creating a plurality of partial hypotheses of the spoken sound of the second speaker in substantially real time while the second speaker is speaking in a conversation with the first speaker;

a second-language resegmentation unit in communication with the second-language automatic speech recognition unit, wherein the second-language resegmentation unit is configured to:

merge at least two of the partial hypotheses received from the second-language automatic speech recognition unit; and

resegment the merged partial hypotheses into a second-language translatable segment in the second language, wherein a segment boundary for the second-language translatable segment is determined based on sound from the first speaker in the first language; and

a second-language machine translation unit in communication with the second-language resegmentation unit, that receives the second-language translatable segment in the second language from the second-language resegmentation unit, wherein the second-language machine translation unit outputs a translation of the spoken sound from the second speaker into the first language based on the received second-language translatable segment.

3. The speech translation system of claim 2 , further comprising:

a first microphone for picking up sound from the first speaker, and wherein output from the first microphone is input to the second-language resegmentation unit; and

a second microphone for picking up sound from the second speaker, and wherein output from the second microphone is input to the first-language resegmentation unit.

4. The speech translation system of claim 3 , wherein the first microphone that picks up sound from the first speaker is in communication with the second-language resegmentation unit via an electronic data communication network.

5. The speech translation system of claim 4 , wherein the second microphone that picks up sound from the second speaker is in communication with the first-language resegmentation unit via the electronic data communication network.

6. The speech translation system of claim 3 , wherein the segment boundary for the first-language translatable segment is determined based recognized speech from the second speaker.

7. The speech translation system of claim 6 , wherein the segment boundary for the second-language translatable segment is determined based recognized speech from the first speaker.

8. The speech translation system of claim 1 , wherein the segment boundary for the first-language translatable segment is determined based on translated speech from the second speaker in the second language.

9. The speech translation system of claim 1 , wherein the segment boundary for the first-language translatable segment is determined based on at least one additional indicator.

10. The speech recognition system of claim 9 , wherein the at least one additional indicator comprises a prosodic cue of the first speaker indicative of the segment boundary for the segment.

11. A speech translation method for translating speech, comprising:

accepting, by a first-language automatic speech recognition unit, spoken sound from a first speaker in a first language and creating a plurality of partial hypotheses of the spoken sound of the first speaker in substantially real time while the first speaker is speaking;

merging, by a first-language resegmentation unit that is in communication with the first-language automatic speech recognition unit, at least two of the partial hypotheses received from the first-language automatic speech recognition unit;

resegmenting, by the first-language resegmentation unit, the merged partial hypotheses into a first-language translatable segment in the first language, wherein a segment boundary for the first-language translatable segment is determined based on sound from a second speaker in a second language; and

outputting, by a first-language machine translation unit that is in communication with the first-language resegmentation unit and that receives the first-language translatable segment in the first-language from the first-language resegmentation unit, a translation of the spoken sound from the first speaker into the second language based on the received first-language translatable segment.

12. The speech translation method of claim 11 , further comprising:

accepting, by a second-language automatic speech recognition unit, spoken sound from the second speaker in the second language and creating a plurality of partial hypotheses of the spoken sound of the second speaker in substantially real time while the second speaker is speaking:

merging, by a second-language resegmentation unit that is in communication with the second-language automatic speech recognition unit, at least two of the partial hypotheses received from the second-language automatic speech recognition unit;

resegmenting, by the second-language resegmentation unit, the merged partial hypotheses into a second-language translatable segment in the second language, wherein a segment boundary for the second-language translatable segment is determined based on sound from the first speaker in the first language; and

outputting, by a second-language machine translation unit that is in communication with the second-language resegmentation unit and that receives the second-language translatable segment in the second language from the second-language resegmentation unit a translation of the spoken sound from the second speaker in lo the first language based on the received second-language translatable segment.

13. The speech translation method of claim 11 , further comprising:

picking up sound from the first speaker by a first microphone, wherein output from the first microphone is input to the second-language resegmentation unit; and

picking up sound from the second speaker by a second microphone, and wherein output from the second microphone is input to the first-language resegmentation unit.

14. The speech translation method of claim 13 , wherein the first microphone that picks up sound from the first speaker is in communication with the second-language resegmentation unit via an electronic data communication network.

15. The speech translation method of claim 14 , wherein the second microphone that picks up sound from the second speaker is in communication with the first-language resegmentation unit via the electronic data communication network.

16. The speech translation method of claim 11 , wherein the segment boundary for the first-language translatable segment is determined based recognized speech from the second speaker.

17. The speech translation method of claim 16 , wherein the segment boundary for the second-language translatable segment is determined based recognized speech from the first speaker.

18. The speech translation method of claim 11 , wherein the segment boundary for the first-language translatable segment is determined based on translated speech from the second speaker in the second language.

19. The speech translation method of claim 11 , wherein the segment boundary for the first-language translatable segment is determined based on at least one additional indicator.

20. The speech recognition system of claim 19 , wherein the at least one additional indicator comprises a prosodic cue of the first speaker indicative of the segment boundary for the segment.