KUDO showcases AI-powered speech-to-speech translation tool at launch event

KUDO在发布会上展示人工智能语音翻译工具

2023-01-24 17:30 multilingual

本文共484个字,阅读需5分钟

阅读模式 切换至中文

On Jan. 24, KUDO held a launch event announcing its forthcoming artificial intelligence (AI)-powered speech-to-speech translation tool, KUDO AI. The tool will soon be available on the New York-based language service provider’s interpreting platform. At first, KUDO’s speech translator will only be available in English, Spanish, French, Portuguese, and German, however the company says that it will update the tool to offer new languages bi-weekly. At the launch event on Tuesday, the company’s leadership conducted a demonstration of KUDO AI. “This is in line with our vision to make every meeting more inclusive, more productive, and truly global,” said the company’s chief language officer and co-founder, Ewandro Magalhães, at the event. The company anticipates that KUDO AI will be available commercially in March. Unlike other speech-to-speech translation tools, KUDO AI will produce simultaneous translations of spoken language, rather than the less seamless sentence-by-sentence translations that other forms of this technology often produce. While KUDO AI will first launch as a pilot program via the KUDO platform, the company notes that it plans to bring it to partner platforms like Microsoft Teams and Zoom. According to a press release announcing KUDO AI, the tool will be equipped with technology that analyzes a speaker’s voice in real-time — as the tool improves, this technology will allow users to customize voices and control speech fluency. “We created a system that analyzes speech structure in real time,” Dr. Claudio Fantinuoli, the company’s head of innovation, said at the launch event. “During this real-time process we reformulate the speech and make it more structured.” At the launch event, KUDO’s leadership acknowledged that the technology isn’t quite perfect — for instance, there can be a bit of a delay if speakers are talking too quickly. At the same time, the company believes KUDO AI has significant potential, with the company’s CEO and co-founder Fardad Zabetian calling it a “paradigm shift … in multilingual communication.” “Let’s give everyone the power to understand, and be understood, in their own language,” Zabetian told those in attendance at the event. Despite its capabilities, KUDO emphasizes the fact that its new technology will never be able to replace human interpreters. Magalhães said that, as an interpreter, he was a bit skeptical of KUDO AI at first, but soon opened up to it as he began using it. In the press release, he added that the tool will provide basic language access services for cases in which a human interpreter might be too costly or impractical to work with. “In many use cases, this is no substitute for human interpreters,” Magalhães said in the press release. “A skilled linguist can read the audience in ways a machine cannot, derive meaning from intonation, accommodate sarcasm, and make adaptations and corrections as (s)he goes along.”
1月24日,KUDO举行发布会,宣布即将推出的人工智能(AI)驱动的语音到语音翻译工具KUDO AI。 该工具将很快在这家总部位于纽约的语言服务提供商的口译平台上提供。起初,KUDO的语音翻译器将只提供英语,西班牙语,法语,葡萄牙语和德语,但该公司表示,它将更新工具,提供新的语言双周。在周二的发布会上,公司领导层进行了KUDO AI的演示。 该公司首席语言官兼联合创始人Ewandro Magalhães在活动中表示:“这符合我们的愿景,即让每一场会议更具包容性、更有成效、真正全球化。” 公司预计KUDO AI将于3月实现商用。与其他语音对语音翻译工具不同,KUDO AI将生成口语的同声翻译,而不是像其他形式的技术那样,生成不那么无缝的逐句翻译。虽然KUDO AI将首先通过KUDO平台作为试点项目推出,但该公司表示,它计划将其带到微软团队和Zoom等合作伙伴平台上。 根据宣布KUDO AI的新闻稿,该工具将配备实时分析说话者声音的技术--随着该工具的改进,这项技术将允许用户定制声音和控制语音流畅度。 “我们创建了一个实时分析语音结构的系统,”该公司创新负责人克劳迪奥·范蒂诺利博士在发布会上说。“在这个实时过程中,我们重新组织了演讲,使其更具结构性。” 在发布会上,KUDO的领导层承认这项技术并不完美--例如,如果演讲者讲得太快,可能会有一点延迟。与此同时,该公司认为KUDO AI具有巨大的潜力,该公司首席执行官兼联合创始人Fardad Zabetian称之为“多语言交流的......范式转变”。 “让我们给每个人的权力,理解,并被理解,在自己的语言,”扎贝蒂安告诉那些出席活动。 尽管它的能力,工藤强调,它的新技术将永远无法取代人类口译员。Magalhães表示,作为一名口译员,他起初对KUDO AI持怀疑态度,但在开始使用后,很快就接受了它。在新闻稿中,他补充说,该工具将为人工口译员可能过于昂贵或不切实际的情况提供基本的语言访问服务。 Magalhães在新闻稿中说:“在许多情况下,这不能替代人类翻译。”“一个熟练的语言学家可以用机器无法做到的方式读懂听众,从语调中获得意义,适应讽刺,并在他说话的时候做出调整和纠正。”

以上中文文本为机器翻译,存在不同程度偏差和错误,请理解并参考英文原文阅读。

阅读原文