今日论文合集：cs.SD语音14篇，eess.AS音频处理20篇。cs.SD语音【1】 Cascaded encoders for fine-tuning ASR models on overlapped speech标题：用于交叠语音ASR模型微调的级联编码器链接：https://arxiv.org/abs/2306.16398作者：Richard Rose,Oscar Chang,Olivier Siohan机构：Google Inc., New York【2】 MC-SpEx: Towards Effective Speaker Extraction with Multi-Scale Interfusion and Conditional Speaker Modulation标题：MC-Spex：基于多尺度融合和条件说话人调制的有效说话人提取链接：https://arxiv.org/abs/2306.16250作者：Jun Chen,Wei Rao,Zilin Wang,Jiuxin Lin,Yukai Ju,Shulin He,Yannan Wang,Zhiyong Wu机构：Shenzhen International Graduate School, Tsinghua University, Shenzhen, China, Tencent Ethereal Audio Lab, Tencent, Shenzhen, China, The Chinese University of Hong Kong, Hong Kong SAR, China备注：Accepted by InterSpeech 2023【3】 Focus on the Sound around You: Monaural Target Speaker Extraction via Distance and Speaker Information标题：聚焦于你周围的声音：基于距离和说话人信息的单声道目标说话人提取链接：https://arxiv.org/abs/2306.16241作者：Jiuxin Lin,Peng Wang,Heinrich Dinkel,Jun Chen,Zhiyong Wu,Zhiyong Yan,Yongqing Wang,Junbo Zhang,Yujun Wang机构：Shenzhen International Graduate School, Tsinghua University, Shenzhen, China, Xiaomi Inc., Beijing, China备注：Accepted by InterSpeech2023【4】 UnitSpeech: Speaker-adaptive Speech Synthesis with Untranscribed Data标题：UnitSpeech：基于未转录数据的说话人自适应语音合成链接：https://arxiv.org/abs/2306.16083作者：Heeseung Kim,Sungwon Kim,Jiheum Yeom,Sungroh Yoon机构：Data Science and AI Lab, ECE, Seoul National University, Seoul , Korea, Interdisciplinary Program in AI, Seoul National University, Seoul , Korea备注：INTERSPEECH 2023, Oral【5】 Graph neural networks for sound source localization on distributed microphone networks标题：基于图神经网络的分布式麦克风网络声源定位链接：https://arxiv.org/abs/2306.16081作者：Eric Grinstein,Mike Brookes,Patrick A. Naylor机构：Department of Electrical and Electronic Engineering, Imperial College London, U.K.备注：Presented as a poster at ICASSP 2023【6】 Improving Primate Sounds Classification using Binary Presorting for Deep Learning标题：基于深度学习的二值预分类改进灵长类语音分类链接：https://arxiv.org/abs/2306.16054作者：Michael Kölle,Steffen Illium,Maximilian Zorn,Jonas Nüßlein,Patrick Suchostawski,Claudia Linnhoff-Popien机构：Institute of Informatics, LMU Munich, Oettingenstraße , Munich, Germany备注：DeLTA【7】 Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound标题：利用对抗性超声对说话人识别系统进行注册阶段的后门攻击链接：https://arxiv.org/abs/2306.16022作者：Xinfeng Li,Junning Ze,Chen Yan,Yushi Cheng,Xiaoyu Ji,Wenyuan Xu【8】 Reprogramming Audio-driven Talking Face Synthesis into Text-driven标题：将音频驱动的对话人脸合成重新编程为文本驱动链接：https://arxiv.org/abs/2306.16003作者：Jeongsoo Choi,Minsu Kim,Se Jin Park,Yong Man Ro机构：School of Electrical Engineering, KAIST, South Korea【9】 Enhanced Neural Beamformer with Spatial Information for Target Speech Extraction标题：基于空间信息的增强型神经波束形成器目标语音提取链接：https://arxiv.org/abs/2306.15942作者：Aoqi Guo,Junnan Wu,Peng Gao,Wenbo Zhu,Qinwen Guo,Dazhi Gao,Yujun Wang机构：Ocean University of China, China, Xiaomi Inc., Beijing, China【10】 Fake the Real: Backdoor Attack on Deep Speech Classification via Voice Conversion标题：伪装真实：通过语音转换对深度语音分类的后门攻击链接：https://arxiv.org/abs/2306.15875作者：Zhe Ye,Terui Mao,Li Dong,Diqun Yan机构：Ningbo University, Ningbo, China, Zhejiang Key Laboratory of Mobile Network Application Technology, China, Ningbo City College of Vocational Technology, Ningbo, China备注：Accepted by INTERSPEECH 2023【11】 Classification of Infant Sleep/Wake States: Cross-Attention among Large Scale Pretrained Transformer Networks using Audio, ECG, and IMU Data标题：婴儿睡眠/清醒状态的分类：使用音频、心电和IMU数据在大规模预先训练的Transformer网络中的交叉注意链接：https://arxiv.org/abs/2306.15808作者：Kai Chieh Chang,Mark Hasegawa-Johnson,Nancy L. McElwain,Bashima Islam机构：∗ Department of Electrical and Computer Engineering, University of Illinois at Urbana Champaign, USA, † Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, USA备注：Preprint for APSIPA2023【12】 Long-term Conversation Analysis: Exploring Utility and Privacy标题：长期对话分析：探索效用和隐私链接：https://arxiv.org/abs/2306.16071作者：Francesco Nespoli,Jule Pohlhausen,Patrick A. Naylor,Joerg Bitzer机构：Microsoft, London, UK, Dept. Electrical and Electronic Engineering, Imperial College London, Institute of Hearing Technology and Audiology, Jade University of Applied Science, Oldenburg, Germany, Fraunhofer IDMT Dept. HSA, Oldenburg, Germany, ∗Equal contribution备注：Submitted to ITG Conference on Speech Communication, 2023【13】 Two-Stage Voice Anonymization for Enhanced Privacy标题：增强隐私的两阶段语音匿名化链接：https://arxiv.org/abs/2306.16069作者：Francesco Nespoli,Daniel Barreda,Joerg Bitzer,Patrick A. Naylor机构：Microsoft, London, UK, Fraunhofer Institute for Digital Media Technology IDMT, Oldenburg, Germany, Imperial College, London, UK备注：submitted to INTERSPEECH【14】 Voicebox: Text-Guided Multilingual Universal Speech Generation at Scale标题：VoiceBox：文本引导的大规模多语言通用语音生成链接：https://arxiv.org/abs/2306.15687作者：Matthew Le,Apoorv Vyas,Bowen Shi,Brian Karrer,Leda Sari,Rashel Moritz,Mary Williamson,Vimal Manohar,Yossi Adi,Jay Mahadeokar,Wei-Ning Hsu机构：Meta AIeess.AS音频处理【1】 Long-term Conversation Analysis: Exploring Utility and Privacy标题：长期对话分析：探索效用和隐私链接：https://arxiv.org/abs/2306.16071作者：Francesco Nespoli,Jule Pohlhausen,Patrick A. Naylor,Joerg Bitzer机构：Microsoft, London, UK, Dept. Electrical and Electronic Engineering, Imperial College London, Institute of Hearing Technology and Audiology, Jade University of Applied Science, Oldenburg, Germany, Fraunhofer IDMT Dept. HSA, Oldenburg, Germany, ∗Equal contribution备注：Submitted to ITG Conference on Speech Communication, 2023【2】 Two-Stage Voice Anonymization for Enhanced Privacy标题：增强隐私的两阶段语音匿名化链接：https://arxiv.org/abs/2306.16069作者：Francesco Nespoli,Daniel Barreda,Joerg Bitzer,Patrick A. Naylor机构：Microsoft, London, UK, Fraunhofer Institute for Digital Media Technology IDMT, Oldenburg, Germany, Imperial College, London, UK备注：submitted to INTERSPEECH【3】 Confidence-based Ensembles of End-to-End Speech Recognition Models标题：基于置信度的端到端语音识别模型集成链接：https://arxiv.org/abs/2306.15824作者：Igor Gitman,Vitaly Lavrukhin,Aleksandr Laptev,Boris Ginsburg机构：NVIDIA, USA, ITMO University, Russia备注：To appear in Proc. INTERSPEECH 2023, August 20-24, 2023, Dublin, Ireland【4】 Scaling Laws for Discriminative Speech Recognition Rescoring Models标题：判别性语音识别研究模型的标度律链接：https://arxiv.org/abs/2306.15815作者：Yile Gu,Prashanth Gurunath Shivakumar,Jari Kolehmainen,Ankur Gandhe,Ariya Rastrow,Ivan Bulyko机构：Amazon Alexa【5】 Voicebox: Text-Guided Multilingual Universal Speech Generation at Scale标题：VoiceBox：文本引导的大规模多语言通用语音生成链接：https://arxiv.org/abs/2306.15687作者：Matthew Le,Apoorv Vyas,Bowen Shi,Brian Karrer,Leda Sari,Rashel Moritz,Mary Williamson,Vimal Manohar,Yossi Adi,Jay Mahadeokar,Wei-Ning Hsu机构：Meta AI【6】 Master-ASR: Achieving Multilingual Scalability and Low-Resource Adaptation in ASR with Modular Learning标题：Master-ASR：通过模块化学习实现ASR中的多语言可伸缩性和低资源适应性链接：https://arxiv.org/abs/2306.15686作者：Zhongzhi Yu,Yang Zhang,Kaizhi Qian,Yonggan Fu,Yingyan Lin机构：IntroductionRecent breakthroughs in deep neural networks (DNNs) havesignificantly advanced the performance of automatic speech 1School of Computer Science, Georgia Institute of Technology【7】 Implementing contextual biasing in GPU decoder for online ASR标题：在线ASR中GPU解码器中上下文偏置的实现链接：https://arxiv.org/abs/2306.15685作者：Iuliia Nigmatulina,Srikanth Madikeri,Esaú Villatoro-Tello,Petr Motliček,Juan Zuluaga-Gomez,Karthik Pandia,Aravind Ganapathiraju机构：Idiap Research Institute, Switzerland ,University of Zurich, Switzerland, Brno University of Technology, Czech Republic, LIDIAP, Ecole Polytechnique Federale de Lausanne (EPFL), Switzerland, Uniphore, India备注：Accepted to Interspeech 2023【8】 Cascaded encoders for fine-tuning ASR models on overlapped speech标题：用于交叠语音ASR模型微调的级联编码器链接：https://arxiv.org/abs/2306.16398作者：Richard Rose,Oscar Chang,Olivier Siohan机构：Google Inc., New York【9】 MC-SpEx: Towards Effective Speaker Extraction with Multi-Scale Interfusion and Conditional Speaker Modulation标题：MC-Spex：基于多尺度融合和条件说话人调制的有效说话人提取链接：https://arxiv.org/abs/2306.16250作者：Jun Chen,Wei Rao,Zilin Wang,Jiuxin Lin,Yukai Ju,Shulin He,Yannan Wang,Zhiyong Wu机构：Shenzhen International Graduate School, Tsinghua University, Shenzhen, China, Tencent Ethereal Audio Lab, Tencent, Shenzhen, China, The Chinese University of Hong Kong, Hong Kong SAR, China备注：Accepted by InterSpeech 2023【10】 Focus on the Sound around You: Monaural Target Speaker Extraction via Distance and Speaker Information标题：聚焦于你周围的声音：基于距离和说话人信息的单声道目标说话人提取链接：https://arxiv.org/abs/2306.16241作者：Jiuxin Lin,Peng Wang,Heinrich Dinkel,Jun Chen,Zhiyong Wu,Zhiyong Yan,Yongqing Wang,Junbo Zhang,Yujun Wang机构：Shenzhen International Graduate School, Tsinghua University, Shenzhen, China, Xiaomi Inc., Beijing, China备注：Accepted by InterSpeech2023【11】 UnitSpeech: Speaker-adaptive Speech Synthesis with Untranscribed Data标题：UnitSpeech：基于未转录数据的说话人自适应语音合成链接：https://arxiv.org/abs/2306.16083作者：Heeseung Kim,Sungwon Kim,Jiheum Yeom,Sungroh Yoon机构：Data Science and AI Lab, ECE, Seoul National University, Seoul , Korea, Interdisciplinary Program in AI, Seoul National University, Seoul , Korea备注：INTERSPEECH 2023, Oral【12】 Graph neural networks for sound source localization on distributed microphone networks标题：基于图神经网络的分布式麦克风网络声源定位链接：https://arxiv.org/abs/2306.16081作者：Eric Grinstein,Mike Brookes,Patrick A. Naylor机构：Department of Electrical and Electronic Engineering, Imperial College London, U.K.备注：Presented as a poster at ICASSP 2023【13】 Improving Primate Sounds Classification using Binary Presorting for Deep Learning标题：基于深度学习的二值预分类改进灵长类语音分类链接：https://arxiv.org/abs/2306.16054作者：Michael Kölle,Steffen Illium,Maximilian Zorn,Jonas Nüßlein,Patrick Suchostawski,Claudia Linnhoff-Popien机构：Institute of Informatics, LMU Munich, Oettingenstraße , Munich, Germany备注：DeLTA【14】 Enrollment-stage Backdoor Attacks on Speaker Recognition Systems via Adversarial Ultrasound标题：利用对抗性超声对说话人识别系统进行注册阶段的后门攻击链接：https://arxiv.org/abs/2306.16022作者：Xinfeng Li,Junning Ze,Chen Yan,Yushi Cheng,Xiaoyu Ji,Wenyuan Xu【15】 Accelerating Transducers through Adjacent Token Merging标题：通过相邻令牌合并加速传感器链接：https://arxiv.org/abs/2306.16009作者：Yuang Li,Yu Wu,Jinyu Li,Shujie Liu机构：University of Cambridge,Microsoft备注：Interspeech 2023【16】 Prompting Large Language Models for Zero-Shot Domain Adaptation in Speech Recognition标题：语音识别中零击域自适应的大语言模型提示链接：https://arxiv.org/abs/2306.16007作者：Yuang Li,Yu Wu,Jinyu Li,Shujie Liu机构：University of Cambridge, Microsoft【17】 Reprogramming Audio-driven Talking Face Synthesis into Text-driven标题：将音频驱动的对话人脸合成重新编程为文本驱动链接：https://arxiv.org/abs/2306.16003作者：Jeongsoo Choi,Minsu Kim,Se Jin Park,Yong Man Ro机构：School of Electrical Engineering, KAIST, South Korea【18】 Enhanced Neural Beamformer with Spatial Information for Target Speech Extraction标题：基于空间信息的增强型神经波束形成器目标语音提取链接：https://arxiv.org/abs/2306.15942作者：Aoqi Guo,Junnan Wu,Peng Gao,Wenbo Zhu,Qinwen Guo,Dazhi Gao,Yujun Wang机构：Ocean University of China, China, Xiaomi Inc., Beijing, China【19】 Fake the Real: Backdoor Attack on Deep Speech Classification via Voice Conversion标题：伪装真实：通过语音转换对深度语音分类的后门攻击链接：https://arxiv.org/abs/2306.15875作者：Zhe Ye,Terui Mao,Li Dong,Diqun Yan机构：Ningbo University, Ningbo, China, Zhejiang Key Laboratory of Mobile Network Application Technology, China, Ningbo City College of Vocational Technology, Ningbo, China备注：Accepted by INTERSPEECH 2023【20】 Classification of Infant Sleep/Wake States: Cross-Attention among Large Scale Pretrained Transformer Networks using Audio, ECG, and IMU Data标题：婴儿睡眠/清醒状态的分类：使用音频、心电和IMU数据在大规模预先训练的Transformer网络中的交叉注意链接：https://arxiv.org/abs/2306.15808作者：Kai Chieh Chang,Mark Hasegawa-Johnson,Nancy L. McElwain,Bashima Islam机构：∗ Department of Electrical and Computer Engineering, University of Illinois at Urbana Champaign, USA, † Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, USA备注：Preprint for APSIPA2023机器翻译，仅供参考
每日语音速递
