The Breakthroughs of Voice AI with Sogou Voice Interaction Technology Centre

与搜狗语音交互中心技术总监陈伟探讨语音人工智能技术的突破和前景  人的感知系统是多方位的，声音是目前在人工智能领域应用较为广泛的主题之一。搜狗在2012年成立语音识别团队，探索智能语音、机器翻译等科技，是中国语音技术的领先者。这一期播客，我们邀请到搜狗语音交互中心的技术总监陈伟，介绍搜狗翻译产品背后的技术突破，包括语音识别、机器翻译和语音合成等。陈伟及其所在的团队还开发了知音os系统，践行着”语音交互+知识计算”战略。这一技术的边缘还延伸到了硬件设施、视觉识别等领域。收听这期播客，还能获得免费票参加NewCo上海创新嘉年华2018，参观商汤科技、哔哩哔哩、IDEO、新车间、唐硕创新体验咨询等公司，与公司高管见面交流。票数有限，先到先得。https://chinaccelerator.com/newco-shanghai-2018/The system of human perceptions is multi-faceted and voice is one of the perceptions that people pay the most attention to in Artificial Intelligence areas. Sogou established a speech recognition team in 2012 to explore technologies such as intelligent voice and machine translation. Now it is obviously a leader in the Chinese voice technology field. In this episode, we’ve invited Chen Wei, the Chief Scientist of Sogou Voice Interaction Technology Centre, to introduce the technological breakthroughs in speech recognition, machine translation and speech synthesis behind Sogou translation products.Chen Wei and his team also developed the Zhiyin OS system which is a system with multi-modal perception capabilities including speech recognition, handwriting recognition and lip reading ability. The edge of this technology also extends to areas such as hardware and visual identification.Perks for our listeners: You can find a promotion code to get free tickets for NewCo Shanghai in this episode. Limited tickets and first come, first serve.https://chinaccelerator.com/newco-shanghai-2018/Show notes:01:23 Introduce Chen Wei02:14 His interest in voice AI technology03:17 Simultaneous translation demo in Chinese, English and Spanish04:56 The job of the Voice Interaction Technology Centre04:49 The technology breakthroughs of the translator06:09 The technology challenges of speech recognition in 42 languages   12:44  How does the Sogou input method work15:10 Explain the meaning of “Zhiyin” in Chinese15:39 The reason that Sogou creates a new system Zhiyin16:40 Realize AI technology in different Sogou products and develop Zhiyin system for different partners18:24 Why Sogou produces a translation hardware device19:37 Other interfaces for translation except for the voice21:21 Future projects, such as lip recognition and virtual anchor24:07 Potential partnerships24:41 The gap between the industry and the universities25:22 How to contact Chen Wei