牛哥精选 · 半年

📋 全部 🤖 AI·大模型 ⚡ 效率工具 📝 深度技术 🚀 产品观察 💰 商业科技 🔓 开源项目 🎨 设计创意 📖 阅读推荐 🏷 资源合集 🌱 成长效率

📝 深度技术 arXiv 计算机视觉 2026-06-30

VIB-AVSR: Variational Information Bottleneck for Noise-Robust LLM-Based Audio-Visual Speech Recognition

突破噪声干扰：变分信息瓶颈让大模型音视频语音识别更稳，INTERSPEECH 2026录用论文。

arXiv:2606.29632v1 Announce Type: cross Abstract: Audio-Visual Speech Recognition takes two input modalities, acoustic and visual streams, where visua…

语音识别音频视觉融合噪声鲁棒性大模型信息瓶颈

📝 深度技术 arXiv NLP 2026-06-09

Is Text All You Need? Text as a Universal Information Bottleneck for Speech LLMs

探讨文本作为信息瓶颈在语音大模型中的核心作用，为连续声学信号集成提供新思路

arXiv:2606.09366v1 Announce Type: new Abstract: Large language models (LLMs) provide a powerful reasoning backbone for speech understanding, but integ…

语音大模型文本信息瓶颈连续声学信号 llm推理语音理解

📝 深度技术 arXiv 机器学习 2026-05-20

Reasoning as Compression: Unifying Budget Forcing via the Conditional Information Bottleneck

通过条件信息瓶颈将推理中的预算强制统一为压缩问题，揭示推理与信息论的内在联系。

arXiv:2603.08462v2 Announce Type: replace Abstract: \ac{CoT} prompting improves LLM accuracy on complex tasks but often increases token usage and infe…

推理压缩信息瓶颈预算强制统一框架

📅 日期

2026-05-20 2026-05-19

🐂 牛哥精选

VIB-AVSR: Variational Information Bottleneck for Noise-Robust LLM-Based Audio-Visual Speech Recognition

Is Text All You Need? Text as a Universal Information Bottleneck for Speech LLMs

Reasoning as Compression: Unifying Budget Forcing via the Conditional Information Bottleneck

📅 日期