论文解读
目标缺席要稳、声纹漂移要跟：流式提取中启发式前沿为何卡住，可学习的说话人状态如何跨过
论文冻结分离主干只换状态机制，用 22 组启发式更新测出稳定性-可塑性前沿，再用闭环元训练的 41k 参数锚定快权重在严重失配下达 10.9 dB 并在 30 s 缺席后保留 6.2 dB，代价是匹配短句略低于静态注册且依赖通道增强覆盖。
元学习
每日研究速递
acl-2026 论文深度解读
共收录 199 篇 acl-2026 会议论文的 Reader 深度解读
论文解读
只给一小时新语言音频：把快速适应做成内外环分工的语音表示学习
论文把新语言低资源表示学习定义为跨语言元学习，用多任务自适应预训练构造适应 эпизод并以一阶双层优化实现可扩展更新，在三语平均上用 1 小时无标注适应追平 6000 小时域内上限，但零样本起点变弱且外环仍依赖源语言音素标注。
论文解读
MetaSICL: Globalizing Auditory LLMs for Underserved Speakers and Languages via Meta Speech In-Context Learning
音频理解 | 8.0/10
每日研究速递
语音/音乐/音频论文速递 2026-08-25
共分析 46 篇语音/AI 论文
论文解读
Normal-Anchored First-Order Model-Agnostic Meta-Learning based Whisper Fine-Tuning for Enhancing Fairness of Cleft Lip and Palate Speech Recognition
语音识别 | 5.9/10
每日研究速递
语音/音乐/音频论文速递 2026-08-04
共分析 39 篇语音/AI 论文
论文解读
Model-Agnostic Meta-Learning Initialization for Distributed Multichannel Active Noise Control
主动降噪 | 5.1/10
论文解读
Versatile On-device Adaptation at the Edge by Unifying Few-shot, Zero-shot, Continual, and In-context Learning
元学习 | 8.1/10
每日研究速递
语音/音乐/音频论文速递 2026-08-03
共分析 16 篇语音/AI 论文
论文解读
Few-Shot Open-Set Audio Classification Using Attention Information-Fused Prototypes
音频分类 | 6.8/10
论文解读
Pmeta-TLA: Backdoor Attacks for Speech Classification Models via Meta-Learning with Timbre Leakage Attack
语音唤醒 | 6/10
每日研究速递
语音/音乐/音频论文速递 2026-07-03
共分析 31 篇语音/AI 论文
论文解读
L-Proto: Language-Aware Episodic Prototypical Training for Multilingual Speaker Verification
说话人验证 | 7.1/10
每日研究速递
语音/音乐/音频论文速递 2026-06-17
共分析 35 篇语音/AI 论文
论文解读
Does language matter for spoken word classification? A multilingual generative meta-learning approach
音频分类 | 6.0/10
论文解读
Scaling few-shot spoken word classification with generative meta-continual learning
音频分类 | 7.0/10
每日研究速递
语音/音乐/音频论文速递 2026-05-14
共分析 16 篇语音/AI 论文
论文解读
A Noval Monte Carlo Gradient Method Based on Meta-Learning for Effective Step-Size Selection in Active Noise Control
噪声控制 | 6.5/10
论文解读
Co-Initialization of Control Filter and Secondary Path via Meta-Learning for Active Noise Control
音频安全 | 7.5/10