解讀智能語音技術的2020：跨語種語音技術成高頻關鍵詞，商業化“加速度”落地

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"italic"},{"type":"strong"}],"text":"本文是 InfoQ“解讀 2020”年終技術盤點系列文章之一。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"2016 年，深度學習和深度神經網絡的突破使得智能語音識別的準確率第一次達到了人類水平，也促使智能語音技術進入到落地階段。尤其是近幾年，語音識別技術逐漸走向成熟，在萬物互聯趨勢下，智能語音技術在教育、金融等各個行業落地日益深入。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"2020年，智能語音賽道的發展態勢如何，有哪些重大的技術突破，在各行業有哪些應用和落地進展？2021年，這個賽道又蘊含着哪些發展機會？InfoQ 採訪了網易有道AI語音團隊的負責人孫豔慶，對智能語音領域過去一年的發展進行總結、回顧與探討，並展望明年的發展趨勢。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"領域湧現多項重大技術突破，挑戰猶在"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"剛剛過去的2020年，智能語音領域出現了多個重要的技術創新與突破。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"“跨語種語音技術是智能語音技術在2020年的高頻關鍵詞。無論是語音識別還是語音合成，都有這樣的形式”，孫豔慶表示。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"例如，谷歌在2019年提出了跨語種的TTS；蘋果在IOS14中，推出了新版翻譯功能，支持自動的語言和語音識別，無需用戶手動預先選擇語言。網易有道在今年9月上線的王源“明星語音”也借鑑了這個框架。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"**孫豔慶認爲，Transformer在更多領域方向（語音\/圖像）、以及建模技術的不斷髮展，使得其建模精度越來越強，也是2020的一個關鍵技術突破。**今年10月，網易有道上線了基於Transformer+CTC 架構的新一代有道ASR引擎，實踐驗證，語音識別準確率和用戶體驗得到了顯著提升，大幅超越了線上採用主流算法的效果。"}]}]}