AI资讯日报 - 2026/6/19

👨‍🔬 Joshua Engels, Callum McDougall, Bilal Chughtai, Janos Kramar, Senthoran Rajamanoharan, Cindy Wu, Arthur Conmy, Asic Q Chen, Jean Tarbouriech, Min Ma, Brendan O'Donoghue, João Gabriel Lopes de Oliveira, Rohin Shah, Neel Nanda

Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages

学术论文 ArXiv 重要度: 9

将代码生成基准扩展到12种语言，揭示LLM在Python之外性能显著下降。

👨‍🔬 Maria Ivanova, Pavel Zadorozhny, Rodion Levichev, Ivan Petrov, Adamenko Pavel, Ivan Lopatin, Alexey Kutalev, Dmitrii Babaev

What Do Safety-Aligned LLMs Learn From Mixed Compliance Demonstrations?

学术论文 ArXiv 重要度: 8

混合示范影响LLM有害合规性，偏好优化是关键，存在近因偏差。

👨‍🔬 Sihui Dai, Mann Patel

Efficient and Sound Probabilistic Verification for AI Agents

学术论文 ArXiv 重要度: 7

提出高效概率验证框架，保证AI代理在不确定性下的策略合规性。

👨‍🔬 Alaia Solko-Breslin, Pramod Kaushik Mudrakarta, Mihai Christodorescu, Somesh Jha, Krishnamurthy Dj Dvijotham

FlowEdit: Associative Memory for Lifelong Pronunciation Adaptation in Flow-Matching TTS

学术论文 ArXiv 重要度: 7

流匹配TTS终身发音适应框架，无需重训，大幅降低专有名词错误率。

👨‍🔬 Harshit Singh, Ayush Pratap Singh, Nityanand Mathur

Structuring and Tokenizing Distributed User Interest Context for Generative Recommendation

学术论文 ArXiv 重要度: 6

提出G2Rec框架，统一图结构和语义标记化，提升生成推荐性能。

👨‍🔬 Ruizhong Qiu, Yinglong Xia, Dongqi Fu, Hanqing Zeng, Ren Chen, Xiangjun Fan, Hong Li, Hong Yan, Hanghang Tong

Toward Calibrated Mixture-of-Experts Under Distribution Shift

学术论文 ArXiv 重要度: 6

研究MoE模型在分布偏移下的校准，提出对抗性重加权方法。

👨‍🔬 Gina Wong, Drew Prinster, Suchi Saria, Rama Chellappa, Anqi Liu

How Do Instructions Shape Speech? Cross-Attention Attribution for Style-Captioned Text-to-Speech

学术论文 ArXiv 重要度: 5

首次将DAAM框架应用于语音扩散模型，分析自然语言对语音风格的影响。

👨‍🔬 Nityanand Mathur, Hamees Sayed, Wasim Madha, Apoorv Singh, Sameer Khurana, Akshat Mandloi, Sudarshan Kamath

LedgerAgent: Structured State for Policy-Adherent Tool-Calling Agents

学术论文 ArXiv 重要度: 5

提出LedgerAgent方法，通过分离状态管理提升工具调用代理的策略合规性。

👨‍🔬 Md Nayem Uddin, Amir Saeidi, Eduardo Blanco, Chitta Baral

DeepSWIP: Quotient-WMC Counterfactuals for Neural Probabilistic Logic Programs

学术论文 ArXiv 重要度: 4

为神经概率逻辑程序提供反事实语义，加速推理并揭示校准问题。

👨‍🔬 Saimun Habib, Vaishak Belle, Fengxiang He

SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm

学术论文 ArXiv 重要度: 4

发布大规模SAR-光学-文本多模态数据集，支持跨模态检索和生成。

👨‍🔬 Solène Debuysère, Nicolas Trouvé, Nathan Letheule, Elise Colin, Georgia Channing

Sovereign Execution Brokers: Enforcing Certificate-Bound Authority in Agentic Control Planes

学术论文 ArXiv 重要度: 3

提出SEB运行时边界，通过证书绑定授权确保代理操作安全。

👨‍🔬 Jun He, Deying Yu

🤖 AI资讯日报

📊 今日趋势总结

MIT Non-AI License

Why Boring Businesses Outlast AI Hype Cycles

Ask HN: Is the rate of progress in AI exponential?

Ask HN: Anyone concerned about NYC Local Law 144?

Ask HN: What's the pain using current AI algorithms?

The AI Crackpot Index

Ask HN: What would you read to learn about 'artificial intelligence'?

Show HN: Startup Raising capital through Book Sales

The Next Bill Gates or Albert Einstein in AI 'Chris Clark' – Yourobot

Common Lisp + Machine Learning Internship at Google (Mountain View, CA)

Bioinformatician

NLP, AI, ML, bots – a passing trend or much more? What's your take on this?

How Transparent is DiffusionGemma?