论文解读
发音纠错不能只给分数：让音频语言模型说出错在哪、怎么练
论文研究朗读式聊天发音训练，用 L2-Arctic-plus 提供词级错误解释与可操作建议，通过两阶段指令微调音频语言模型，在误读检测与建议生成上超过级联 ASR 加 LLM 和现有开源模型，代价是依赖合成标注与朗读场景。
指令微调
论文解读
耦合失真下不换模型而换管线：StrixAE 用智能体调度专家模型
针对噪声混响多人交叠耦合且需求因人而异的问题，StrixAE 用多模态大模型做控制器调度专家工具链，经 AcoustBench 思维链微调与音频感知强化学习后，在真实盲测多项感知指标上超过多数开源基线，但感知质量优化仍慢且依赖外部工具。
论文解读
从一次性生成到先诊断再重做：LoopTTS 如何让语音大模型当韵律医生
语音合成 | 7.0/10
论文解读
Do Spoken Language Models Hear Speech as They Read Text? Bridging Structural Gaps Between Speech and Text
语音交互 | 10.0/10
论文解读
SwanTale: Unified Multi-Speaker Speech and Audio Generation for Instruct and Zero-Shot Tasks
音频生成 | 7.1/10
论文解读
Improving Text-to-Audio Instruction Following via Fine-Grained Feedback from Audio-Aware Large Language Models
音频生成 | 7.2/10
论文解读
NAVER LABS System Re-implementation for the IWSLT 2026 Instruction-Following Task
语音翻译 | 6.4/10
论文解读
FutureOmni: Evaluating Future Forecasting from Omni-Modal Context for Multimodal LLMs
音视频问答 | 8/10
论文解读
SonicMaster: Towards Controllable All-in-One Music Restoration and Mastering
音频修复 | 8/10
论文解读
Unlocking Speech–Text Compositional Powers: Instruction-Following Speech Language Models without Instruction Tuning
语音交互 | 6.7/10
论文解读
Aligning MusicLLM with Emotion using Instruction Tuning and Feedback-Driven Alignment
音乐情感识别 | 4.9/10
论文解读
OmniVideo-100K: A Dataset for Audio-Visual Reasoning through Structured Scripts and Evidence Chains
数据增强 | 8.2/10
论文解读
TinyGiantALM: A Compact Audio-Language Model for Intent-Aware Reasoning under Resource Constraints
多模态模型 | 6.4/10
论文解读
SMILE-Next: Teaching Large Language Models to Detect, Classify, and Reason about Laughter
语音情感识别 | 8.7/10
论文解读
TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis
语音质量评估 | 7.5/10
论文解读
Cross-Lingual Alzheimer’s Disease Detection with Multimodal LLMs via Speech Cue-Augmented Prompting and Instruction Tuning
语音生物标志物 | 6.5/10
论文解读
MIDI-LLaMA: An Instruction-Following Multimodal LLM for Symbolic Music Understanding
音乐理解 | 7.5/10
论文解读
TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis
语音合成评估 | 7.0/10
论文解读
TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis
语音质量评估 | 7.5/10