Once the stuff of science fiction (đề tài khoa học viễn tưởng /stʌf əv ˈsaɪəns ˈfɪkʃən/), technology (công nghệ /tɛkˈnɒlədʒi/) that enables people to talk using different
languages (ngụn ngữ /ˈlổŋɡwɪdʒɪz/) is now here. But how effective (hiệu quả /ɪˈfɛktɪv/) is
it?
1.
Noise (tiếng ồn /nɔɪz/), Alex Waibel tells me, is one of the major challenges (thách thức
/ˈʧổlɪndʒɪz/) that artificial speech translation (dịch giọng núi nhõn tạo /ˌɑːrtɪˈfɪʃl spiːʧ trổnsˈleɪʃən/) has to meet. A device (thiết bị /dɪˈvaɪs/) may be able to recognize speech (nhận diện giọng nói /ˈrekəɡnaɪz spiːʧ/) in a laboratory, or a meeting room, but will struggle (gặp khó khăn /ˈstrʌɡəl/) to cope with the kind of background noise (tiếng ồn nền /ˈbổkɡraʊnd nɔɪz/) I can hear in my office surrounding Professor Waibel as he speaks to
me from Kyoto station in Japan. I’m struggling to follow him in English, on a scratchy line
(đường truyền kộm /ˈskrổʧi laɪn/) that reminds me we are nearly 10,000 kilometers apart –
and that distance (khoảng cách /ˈdɪstəns/) is still an obstacle (trở ngại /ˈɒbstəkəl/) to
communication (giao tiếp /kəˌmjuːnɪˈkeɪʃən/) even if you’re speaking the same language,
as we are. We haven’t reached the future (đạt tới tương lai /riːʧt ðə ˈfjuːʧər/) yet. If we had, Waibel would have been able to speak more comfortably in his native German and I would have been able to hear his words in English.
2.
At Karlsruhe Institute of Technology, where he is a professor of computer science, Waibel and his colleagues already give lectures (bài giảng /ˈlɛktʃərz/) in German that their students can follow in English via an electronic translator (máy dịch điện tử /ɪˌlɛkˈtrɒnɪk ˈtrổnzleɪtər/). The system generates text that students can read on their laptops or phones,
so the process is somewhat similar to subtitling (việc tạo phụ đề /ˈsʌbˌtaɪtlɪŋ/). It helps
that lecturers (giảng viên /ˈlɛktʃərz/) speak clearly, don’t have to compete (cạnh tranh /kəmˈpiːt/) with background chatter, and say much the same thing each year.
3.
35 The idea of artificial speech translation has been around for a long time. Douglas Adams’
science fiction novel, The Hitchhiker’s Guide to the Galaxy, published in 1979, featured a
life form (dạng sống /laɪf fɔːrm/) called the ‘Babel fish’ which, when placed in the ear,
enabled a listener (người nghe /ˈlɪsənər/) to understand any language (hiểu bất kỳ ngôn
ngữ nào /ˌʌndərˈstổnd ˈɛni ˈlổŋɡwɪdʒ/) in the universe *(vũ trụ /ˈjuːnɪvɜːrs/). It came to
represent one of those devices that technology enthusiasts dream of long before they become practically realizable, like TVs flat enough to hang on walls: objects (vật thể /ˈɒbdʒɛkts/) that we once could only dream of having but that are now commonplace
*(phổ biến /ˈkɒmənpleɪs/). Now devices that look like prototype Babel fish (nguyên mẫu
cá Babel /ˈprəʊtəʊtaɪp ˈbeɪbəl fɪʃ/) have started to appear, riding a wave of advances (tiến bộ /ədˈvɑːnsɪz/) in artificial translation and voice recognition *(nhận diện giọng nói /vɔɪs
ˌrekəɡˈnɪʃən/).
4.
At this stage (giai đoạn /steɪdʒ/), however, they seem to be regarded (được coi như /rɪˈɡɑːrdɪd/) as eye-catching novelties (đồ mới lạ bắt mắt /aɪ-ˈkổʧɪŋ ˈnɒvəltiz/) rather
than steps towards what Waibel calls ‘making a language-transparent society’ (tạo ra một
xó hội minh bạch ngụn ngữ /ˈlổŋɡwɪdʒ trổnˈspổrənt səˈsaɪəti/). They tend to be domestic devices (thiết bị gia dụng /dəˈmɛstɪk dɪˈvaɪsɪz/) or applications suitable for hotel check- ins *(phù hợp cho việc nhận phòng khách sạn /ˈsuːtəbl fər hoʊˈtɛl ˈʧɛk-ɪnz/), for example,
providing a practical alternative (sự thay thế thực tế /ɔːlˈtɜːrnətɪv/) to speaking traveler’s
nglish *(tiếng Anh cho người du lịch /ˈtrổvələrz ˈɪŋɡlɪʃ/). The efficiency (hiệu quả /ɪˈfɪʃənsi/) of the translator (mỏy dịch /trổnsˈleɪtər/) is less important than the social function *(chức năng xã hội /ˈsoʊʃl ˈfʌŋkʃən/). However, ‘Professionals are less inclined to
be patient (ít có xu hướng kiên nhẫn /ɪnˈklaɪnd tə bi ˈpeɪʃənt/) in a conversation *(cuộc trò
chuyện /ˌkɒnvərˈseɪʃən/),’ founder and CEO at Waverly Labs, Andrew Ochoa, observes
*(nhận xét /əbˈzɜːrvz/). To redress (khắc phục /rɪˈdrɛs/) this, Waverly is now preparing a
new model for professional applications, which entails (bao gồm /ɪnˈteɪlz/) performance
improvements (cải thiện hiệu suất /pərˈfɔːrməns ɪmˈpruːvmənts/) in speech recognition
*(nhận diện giọng nói /spiːʧ ˌrɛkəɡˈnɪʃən/), translation accuracy (độ chính xác của bản dịch
36
/trổnsˈleɪʃən ˈổkjərəsi/) and the time it takes to deliver (cung cấp /dɪˈlɪvər/) the translated
speech.
5.
For a conversation *(cuộc trò chuyện /ˌkɒnvərˈseɪʃən/), both speakers need to have
devices called Pilots (Pilots, một loại tai nghe phiên dịch /ˈpaɪləts/) (translator earpieces) in their ears. ‘We find that there’s a barrier (rào cản /ˈbổriər/) with sharing (chia sẻ /ˈʃɛrɪŋ/)
one of the earphones with a stranger,’ says Ochoa. That can’t have been totally unexpected
*(không thể hoàn toàn bất ngờ /ˌʌnɪkˈspɛktɪd/). The problem would be solved if earpiece translators became sufficiently prevalent (phổ biến /ˈprɛvələnt/) that strangers would be
likely to already have their own in their ears. Whether that happens, and how quickly, will probably depend not so much on the earpieces themselves, but on the prevalence (mức
độ phổ biến /ˈprɛvələns/) of voice-controlled devices (thiết bị điều khiển bằng giọng nói /vɔɪs-kənˈtroʊld dɪˈvaɪsɪz/) and artificial translation (dịch thuật nhân tạo /ˌɑːrtɪˈfɪʃəl trổnsˈleɪʃən/) in general.
6.
Waibel highlights (nhấn mạnh /ˈhaɪlaɪts/) the significance (tầm quan trọng /sɪɡˈnɪfɪkəns/)
of certain Asian nations, noting that voice translation (dịch giọng núi /vɔɪs trổnsˈleɪʃən/)
has really taken off in countries such as Japan with a range of systems *(hệ thống
/ˈsɪstəmz/). There is still a long way to go, though. A translation system (hệ thống dịch /trổnsˈleɪʃən ˈsɪstəm/) needs to be simultaneous *(đồng thời /ˌsaɪmlˈteɪniəs/), like the translator’s voice speaking over the foreign politician (chính trị gia nước ngoài /ˈfɔːrən ˌpɒlɪˈtɪʃən/) being interviewed on the TV, rather than in sections that oblige (buộc /əˈblaɪdʒ/) speakers to pause (tạm dừng /pɔːz/) after every few remarks and wait for the
translation to be delivered *(được cung cấp /dɪˈlɪvərd/). It needs to work offline *(ngoại
tuyến /ˌɒfˈlaɪn/), for situations where internet access (truy cập internet /ˈɪntərˌnɛt ˈổkses/) isn’t possible, and to address apprehensions (lo ngại /ˌổprɪˈhɛnʃənz/) about the
amount of private speech data (dữ liệu giọng nói riêng tư /ˈpraɪvɪt spiːʧ ˈdeɪtə/)
accumulating in the cloud *(đám mây /klaʊd/), having been sent to servers (máy chủ /ˈsɜːrvərz/) for processing *(xử lý /ˈprɑːsesɪŋ/).
7.
37
ystems (hệ thống /ˈsɪstəmz/) not only need to cope with (đối phó với /koʊp wɪð/)
physical challenges (những thỏch thức vật lý /ˈfɪzɪkl ˈʧổlɪndʒɪz/) such as noise *(tiếng ồn
/nɔɪz/), they will also need to be socially aware (nhận thức xã hội /ˈsoʊʃəli əˈweər/) by
addressing people in the right way. Some cultural traditions (truyền thống văn hóa /ˈkʌltʃərəl trəˈdɪʃənz/) demand solemn respect (sự tôn trọng trang nghiêm /ˈsɒləm rɪˈspɛkt/) for academic status *(địa vị học thuật /ˌổkəˈdɛmɪk ˈstổtəs/), for example, and it is only polite (lịch sự /pəˈlaɪt/) to respect this. tiquette-sensitive artificial translators
(mỏy dịch nhạy cảm với nghi thức /ˈɛtɪkɛt-ˈsɛnsətɪv ˌɑːrtɪˈfɪʃəl trổnsˈleɪtərz/) could relieve
people of the need to know these differing cultural norms *(chuẩn mực /nɔːrmz/). At the same time, they might help to preserve local customs *(bảo tồn phong tục địa phương
/prɪˈzɜːrv ˈloʊkəl ˈkʌstəmz/), slowing the spread of habits associated with international nglish *(tiếng Anh quốc tế /ˌɪntərˈnổʃənəl ˈɪŋɡlɪʃ/), such as its readiness to get on first- name terms *(dùng tên riêng /fɜːrst-neɪm tɜːrmz/).
8.
Professors and other professionals (giáo sư và các chuyên gia khác /prəˈfɛsərz ənd ˈʌðər
prəˈfɛʃənəlz/) will not outsource (thuê ngoài, giao phó/ˈaʊtsɔːrs/) language awareness
(nhận thức ngụn ngữ /ˈlổŋɡwɪdʒ əˈwɛrnəs/) to software, though. If the technology matures (trưởng thành /məˈtjʊərz/) into seamless, ubiquitous artificial speech translation *(dịch
giọng nói nhân tạo liền mạch, phổ biến /ˈsiːmləs, juːˈbɪkwɪtəs ˌɑːrtɪˈfɪʃəl spiːʧ trổnsˈleɪʃən/), it will actually add value (gia tăng giỏ trị /ổd ˈvổljuː/) to language skills.
Whether it will help people conduct (thực hiện /kənˈdʌkt/) their family lives (cuộc sống gia
đỡnh /ˈfổmɪli laɪvz/) or relationships is open to question—though one noteworthy possibility (khả năng đáng chú ý /ˈnoʊtwɜːrði ˌpɒsəˈbɪləti/) is that it could overcome the
language barriers (vượt qua rào cản ngụn ngữ /ˌoʊvərˈkʌm ðə ˈlổŋɡwɪdʒ ˈbổriərz/) that
often arise between generations (thế hệ /ˌdʒɛnəˈreɪʃənz/) after migration *(di cư
/maɪˈɡreɪʃən/), leaving children and their grandparents without a shared language *(ngôn ngữ chung /ʃɛrd ˈlổŋɡwɪdʒ/).
9.
Whatever uses it is put to, though, it will never be as good as the real thing. Even if voice- morphing technology (công nghệ biến đổi giọng nói /vɔɪs-ˈmɔːrfɪŋ tɛkˈnɒlədʒi/) simulates
38
(mô phỏng /ˈsɪmjəˌleɪts/) the speaker’s voice, their lip movements (cử động môi /lɪp ˈmuːvmənts/) won’t match, and they will look like they are in a dubbed movie *(phim lồng
tiếng /dʌbd ˈmuːvi/). The contrast (sự tương phản /ˈkɒntrổst/) will underline (nhấn mạnh /ˌʌndərˈlaɪn/) the value of shared languages (giỏ trị của ngụn ngữ chung /ˈvổljuː əv ʃɛrd
ˈlổŋɡwɪdʒɪz/), and the value of learning them. Sharing a language can promote (thỳc đẩy
/prəˈmoʊt/) a sense of belonging (thuộc về /bɪˈlɒŋɪŋ/) and community *(cộng đồng
/kəˈmjuːnɪti/), as with the international scientists who use English as a lingua franca *(ngôn ngữ chung /ˌlɪŋɡwə ˈfrổŋkə/), where their predecessors (người tiền nhiệm /ˈprɛdəsɛsərz/) used Latin. Though the practical need for a common language (ngôn ngữ chung /ˈkɒmən ˈlổŋɡwɪdʒ/) will diminish, the social value (giỏ trị xó hội /ˈsəʊʃəl ˈvổljuː/) of sharing one
will persist (tồn tại /pəˈsɪst/). And software will never be a substitute (thay thế /ˈsʌbstɪtjuːt/) for the subtle but vital understanding (sự hiểu biết tinh tế nhưng quan trọng /ˈsʌtl bʌt ˈvaɪtl ˌʌndərˈstổndɪŋ/) that comes with knowledge of a language.
TEST4