PolyAI Releases Dialog-RSN-1: An Audio-Native Dialog Model That Fuses Turn-Taking, Speech Recognition, Function Calling, And Response
https://www.marktechpost.com/2026/07/30/polyai-releases-dialog-rsn-1-an-audio-native-dialog-model-that-fuses-turn-taking-speech-recognition-function-calling-and-response/📌 【PolyAI 新技術】Dialog-RSN-1 登場:直接「聽」聲音而非讀逐字稿的對話模型
TL;DR:PolyAI 推出音訊原生對話模型 Dialog-RSN-1,將語音辨識與對話生成整合,解決傳統串聯架構丟失語氣資訊的問題。
🎣 傳統的語音助手在對話時,往往像是「聽完、轉成文字、再讀文字」的過程,這導致對話中細微的語氣、猶豫與情緒完全消失。PolyAI 提出的 Dialog-RSN-1 正試圖打破這個僵局。
🤔 傳統架構的權衡與困境
目前主流的語音對話架構主要分為兩大類,但各有其致命傷:
- 串聯式堆疊 (Cascaded Stack):將 ASR(語音辨識)的最佳預測結果傳給 LLM。缺點是語氣、猶豫與辨識不確定性在進入 LLM 前就已消失,且需要手動調整 endpointing 與 ASR 偏置參數,難以跨場景通用。
- 語音對語音模型 (Speech-to-speech):如 GPT Realtime 或 Gemini Live。雖然保留了音訊,但將聲音直接編碼進模型,限制了對發音的控制,且全雙工 (full-duplex) 版本會導致整通電話都佔用 GPU 資源。
🧩 Dialog-RSN-1 的設計理念:音訊感知但非全音訊生成
Dialog-RSN-1 採取了一種折衷但高效的架構:它在輸入端是「音訊感知 (audio-aware)」的,模型直接對原始音訊進行推理,但將生成端交給獨立且可透過提示詞 (promptable) 控制的 TTS 系統。
其運作流程如下:
- 使用高召回率的 VAD (語音活動檢測) 搭配計時器來決定何時執行模型。
- 模型接收原始音訊並進行推理,並根據模型產生的第一個 token 來決定代理人 (agent) 是否應該說話。
- 決定對話輪次 (turn-taking) 是由具備完整上下文的模型做出,而非僅靠廉價的聲學線索。
- 轉錄 (Transcription) 流程會在回應產生後才執行。
📊 針對極低延遲進行的效能最佳化
為了在 A100 GPU 上達成低於 300ms 的延遲目標,PolyAI 針對候選模型(預估在 8B 密集模型至 30B 稀疏模型之間)進行了一系列工程實踐:
- 預填 (Prefilling):在使用者說話時同步預填 attention cache。
- 提示詞優化:使用 append-only 的提示詞模板以減少 cache 失效。
- 路由策略:將每位來電者路由至同一臺 GPU。
- 投機採樣 (Speculative Drafting):使用經過微調的投機草稿模型,平均接受度為 3.9 個 tokens。
- 自動推理:在 RFT (Reinforcement Fine-Tuning) 過程中學習自動推理能力。
⚠️ 目前僅限現有客戶使用
雖然 Dialog-RSN-1 已應用於實際生產環境,但目前並未開放權重或公開 API。現有客戶可以直接啟用,新客戶則需申請早期存取權。
🎯 實務啟示
對於開發語音 AI 的工程師來說,Dialog-RSN-1 證明瞭「音訊感知輸入 + 獨立 TTS 輸出」可能是平衡「語氣理解」與「發音控制」與「計算成本」的最佳實踐路徑。
🔗 來源
- 標題:PolyAI Releases Dialog-RSN-1: An Audio-Native Dialog Model That Fuses Turn-Taking, Speech Recognition, Function Calling, And Response
- 作者/機構:Michal Sutter @ MarkTechPost
- 連結:https://www.marktechpost.com/2026/07/30/polyai-releases-dialog-rsn-1-an-audio-native-dialog-model-that-fuses-turn-taking-speech-recognition-function-calling-and-response/
#AI #MachineLearning #PolyAI #SpeechRecognition #LLM #AudioNative #ConversationalAI #NLP #DeepLearning #TechNews
原始資料 MarkTechPost · 收集於 2026-07-31
摘要原文
PolyAI has introduced Dialog-RSN-1 , a dialog model that perceives the caller’s audio directly instead of reading a transcript. It fuses turn-taking, speech recognition, function calling and response generation into one audio-native model, and is already handling live production calls. Yes, but only through PolyAI: no open weights, no public API yet. Existing customers can enable it today; new customers can request early access . Two architectures dominate, and each concedes something. A cascaded stack sends only the ASR’s best guess to the LLM, so tone, hesitation and recognition uncertainty are gone before the LLM sees anything. Tuning means hand-adjusting end-pointing parameters and ASR biasing that rarely generalize across use cases. Speech-to-speech models such as GPT Realtime and Gemini Live keep the audio but bake the voice into the model, limiting pronunciation control, and always-on full-duplex variants pin a GPU for the entire call. Dialog-RSN-1 is audio-aware on input only: one model reasons over raw audio and hands generation to a separate, promptable TTS system. It is probed on demand rather than streamed: a high-recall VAD plus a few timers decide when to run it, and the first token of the reply settles whether the agent should speak. Cheap acoustic cues only choose when to ask; the model, with full context, makes the actual turn-taking call. PolyAI post-trained open-weight multimodal models with supervised and reinforcement finetuning on in-house data. The pipeline is broadly base-model agnostic; PolyAI evaluated Gemma, GPT-OSS, Qwen and Mistral. Targeting sub-300ms on A100 GPUs puts candidates in the 8B dense to 30B sparse range. Latency work includes prefilling the attention cache while the user speaks, an append-only prompt template to minimize cache invalidation, routing each caller to the same GPU, a finetuned speculative drafter with mean acceptance of 3.9 tokens, and auto-reasoning learned during RFT. Transcription runs last, after the response
由 tencent/hy3:free 自動生成