Hugging Face Blog发布于 08/21 08:00
Measuring benchmark optimization in speech recognition
（翻译）衡量语音识别中的基准优化
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
AWS Machine Learning Blog发布于 08/25 00:13
Building a restaurant telephony AI host with Amazon Connect
（翻译）使用 Amazon Connect 构建餐厅电话 AI 接待员
Learn how to build a voice ordering system for restaurants that answers a phone call and takes an order end to end, with no app, no website, and no sign-in. It uses Amazon Connect for telephony, Amazon Connect Agentic Voice for real-time speech, an Amazon Connect AI agent for reasoning, and Amazon B
Hugging Face Blog发布于 08/11 00:25
Build Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS
（翻译）构建低延迟多语言语音智能体：借助 NVIDIA Magpie TTS 实现开放权重与完整部署控制
A Blog post by NVIDIA on Hugging Face
Google DeepMind Blog发布于 06/09 23:16
Fluid, natural voice translation with Gemini 3.5 Live Translate
（翻译）Gemini 3.5 Live Translate：流畅自然的语音翻译
Gemini 3.5 Live Translate brings near real-time, natural speech translation to Google AI Studio, Google Translate and Google Meet.
Hugging Face Blog发布于 08/25 23:01
Extremely Fast and Accurate Transcription with Granite Speech 5.0 Turbo CTC
（翻译）使用 Granite Speech 5.0 Turbo CTC 实现极其快速且准确的转录
A Blog post by IBM Granite on Hugging Face
Hugging Face Blog发布于 08/28 08:00
The Open ASR Leaderboard Adds Its First Global South Language
（翻译）Open ASR 排行榜新增首个全球南方语言
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
雷峰网发布于 09/04 09:19
从无线麦克风到AI语音入口：猛玛 LARK A2 原生联动 WorkBuddy
9月2日,腾讯 WorkBuddy 生态发布会上宣布,与专业音视频硬件品牌HOLLYLAND猛玛达成深度生态合作,推出猛玛LARK A2无线麦克风WorkBuddy联名款。作为首款原生联动WorkBuddy智能体的AI语音入口硬件,该产品在专业收音基础上,通过机身按键联动WorkBuddy语音输入、内容删除与指令发送,为电脑端AI语音交互提供新的操作方式。
arXiv cs.AI发布于 09/04 12:00
DuplexSpeechBench-IFEval: Evaluating Implicit Instruction Following in Full-Duplex Voice Agents
（翻译）DuplexSpeechBench-IFEval：评估全双工语音智能体中的隐式指令遵循
Abstract page for arXiv paper 2609.03423: DuplexSpeechBench-IFEval: Evaluating Implicit Instruction Following in Full-Duplex Voice Agents
InfoQ 中文发布于 09/08 21:32
OpenAI 详解 GPT-Live 架构如何实现了连续的有状态语音交互
最近，OpenAI 发布了一篇关于 GPT-Live 的工程报告。该报告详细描述了他们如何设计这个系统，将对延迟敏感的媒体处理与更广泛的应用工作分离，同时保持语音交互的连续性。
arXiv cs.AI发布于 09/10 12:00
RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems
（翻译）RESCUE-BENCH：面向关系感知的多方情感支持对话系统
Abstract page for arXiv paper 2609.09657: RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems
arXiv cs.AI发布于 09/12 12:00
Same Day, Same Story; One Day Ahead, a Different Signal: The Dual Validity of Financial Sentiment
（翻译）同日同样叙事；提前一天，信号不同：金融情感分析的双重效度
Abstract page for arXiv paper 2609.11144: Same Day, Same Story; One Day Ahead, a Different Signal: The Dual Validity of Financial Sentiment
arXiv cs.AI发布于 09/14 12:00
Can LLMs in Draft-Verify-Revise Pipelines Resolve Deictic Ambiguity?
（翻译）LLM在草稿-验证-修订流水线中能否解决指示语歧义？
Abstract page for arXiv paper 2609.12162: Can LLMs in Draft-Verify-Revise Pipelines Resolve Deictic Ambiguity?
雷峰网发布于 09/15 17:50
阶跃发布 StepAudio 3 ，多款语音模型登顶 Artificial Analysis 全球榜单
9月15日，阶跃发布新一代语音大模型 StepAudio 3 系列，一次性推出 StepAudio 3 Realtime、StepAudio 3 ASR、StepAudio 3 TTS、StepAudio 3 Gen 和 StepAudio 3 Music 五款模型，覆盖实时语音交互、语音识别、真人级语音生成和音乐创作等场景。其中，多款模型在 权威 第三方 AI 模型评测机构 Artificial Analysis 榜单中位列全球第一。
Google AI Blog发布于 09/16 00:00
AI for everyone in every language
（翻译）让 AI 覆盖每一种语言和每一个人
We’re moving beyond traditional text translation to build models that understand the world’s rich, living languages exactly as they are expressed.
