论文解读
把级联状态机装进大模型：TurnFSM 如何串行做提交与拒绝
TurnFSM 把语义 VAD 与 utterance 级拒绝写成先提交后接受或拒绝的有限状态转移并用一阶依赖实现流式预测，在内部语义 VAD 上报告成功率 82.41% 与成功延迟 80.9 ms，在拒绝上报告 FAR 3.35% 与 FRR 16.04%，代价是仍需两阶段对齐与大模型推理开销。
状态空间模型
论文解读
MambaVoiceCloning: Efficient and Expressive Text-to-Speech via State-Space Modeling and Diffusion Control
语音合成 | 6.5/10
论文解读
MambaVoiceCloning: Efficient and Expressive Text-to-Speech via State-Space Modeling and Diffusion Control
语音克隆 | 7.5/10
论文解读
A Hybrid Convolution-Mamba Network with Tone-Octave Contrastive Learning for Stratified Semi-Supervised Singing Melody Extraction
歌唱旋律提取 | 7.5/10
论文解读
Combining SSL Speech Features, Contextual Transformers and Mamba Models for Realistic Audio Spoofing Detection
音频深度伪造检测 | 7.5/10
论文解读
Condition-Invariant fMRI decoding of speech intelligibility with deep state space model
神经解码 | 7.0/10
论文解读
DECAF: Dynamic Envelope Context-Aware Fusion for Speech-Envelope Reconstruction from EEG
语音增强 | 7.0/10
论文解读
Disentangling Physiology from Fidelity: Latent-Guided Diffusion Models for Cross-Modal Cardiac Synthesis
音频生成 | 7.5/10
论文解读
Mambaformer: State-Space Augmented Self-Attention with Downup Sampling for Monaural Speech Enhancement
语音增强 | 7.0/10
论文解读
Neuromamba: Adaptive Frequency Filtering with a Pyramid Mamba for sEEG-driven Speech Synthesis
语音合成 | 8.0/10
论文解读
Respire-Mamba C-UNet: Consistency-Trained Autoencoder for High-Fidelity Respiratory Sound Compression
音频压缩 | 7.0/10
论文解读
Spring Reverb Emulation with Hybrid Gated Convolutional Networks and State Space Models
音频生成 | 7.5/10
论文解读
Tri-Attention Fusion: Joint Temporal-Spectral and Bidirectional Modeling for Speech Spoofing Detection
语音伪造检测 | 7.0/10
论文解读
HCFD: A Benchmark for Audio Deepfake Detection in Healthcare
本文针对医疗健康领域中神经音频编解码器生成的语音深伪检测问题，提出了一个全新的研究任务（HCFD）和基准数据集（HCFK）。研究发现，在健康语音上训练的现有深伪检测模型在病态语音上性能显著下降。为此，论文首先验证了预训练音频模型（如PaSST）能更好地应对病理语音带来的变异性。更重要的是，本文提出了