AI资讯日报 - 2026/6/30

👨‍🔬 Yen-Jen Wang, Jiaman Li, Sirui Chen, Takara E. Truong, Pei Xu, Pieter Abbeel, Rocky Duan, Koushil Sreenath, Angjoo Kanazawa, Carmelo Sferrazza, Guanya Shi, Karen Liu

C$^{2}$R: Cross-sample Consistency Regularization Mitigates Feature Splitting and Absorption in Sparse Autoencoders

学术论文 ArXiv 重要度: 8

通过跨样本一致性正则化缓解稀疏自编码器的特征分裂与吸收问题，提升可解释性。

👨‍🔬 Haoran Jin, Xiting Wang, Shijie Ren, Hong Xie, Defu Lian

MESA: Prioritizing Vulnerable Communication Channels for Securing Multi-Agent Systems

学术论文 ArXiv 重要度: 8

提出无标签框架MESA，主动识别多智能体系统中关键通信边，优先防御。

👨‍🔬 Kunyang Li, Kyle Domico, Jonathan Gregory, Patrick McDaniel

GROW$^2$: Grounding Which and Where for Robot Tool Use

学术论文 ArXiv 重要度: 7

利用物体部分作为抽象，分层实现开放世界工具选择与操作区域定位，零样本泛化到新物体。

👨‍🔬 Yuhong Deng, Yuyao Liu, David Hsu

DOPD: Dual On-policy Distillation

学术论文 ArXiv 重要度: 7

提出双在线策略蒸馏，通过优势感知动态路由缓解特权幻觉，提升知识蒸馏效果。

👨‍🔬 Xinlei Yu, Gen Li, Qingyi Si, Guibin Zhang, Yuqi Xu, Congcong Wang, Shuai Dong, Kaiwen Tuo, Xiangyu Zeng, Kaituo Feng, Qunzhong Wang, Yang Shi, Xiaobin Hu, Xiangyu Yue, Jiaqi Wang, Shuicheng Yan

Beyond 2D Matching: A Unified Single-Stage Framework for Geometry-Aware Cross-View Object Geo-Localization

学术论文 ArXiv 重要度: 7

提出单阶段几何感知框架GAGeo，结合3D基础模型实现跨视角目标地理定位，显著优于现有方法。

👨‍🔬 Liyao Wang, Ruipu Wu, Haojun Xu, Lei Shi, Linjiang Huang, Si Liu

LeVo 2: Stable and Melodious Song Generation via Hierarchical Representation Modeling and Progressive Post-Training

学术论文 ArXiv 重要度: 6

提出混合LLM-扩散框架，通过分层建模和渐进后训练生成高质量全长歌曲，优于开源基线。

👨‍🔬 Shun Lei, Huaicheng Zhang, Dapeng Wu, Yaoxun Xu, Lishi Zuo, Wei Tan, Hangting Chen, Guangzheng Li, Jianwei Yu, Zhiyong Wu, Dong Yu

A Multi-task Mixture of Experts Framework for Malware Classification, Packing Detection, and Family Attribution

学术论文 ArXiv 重要度: 6

提出多门控专家混合框架，统一完成恶意软件分类、加壳检测和家族归因，鲁棒性高。

👨‍🔬 Jithin S., Roshin Sleeba C., Anvin Mariya P. B., Asmitha K. A., Vinod P., Serena Nicolazzo, Antonino Nocera

Optimization Dynamics Imprint Semantic Specificity in Contrastive Embedding Norms

学术论文 ArXiv 重要度: 5

理论解释对比嵌入中向量长度编码语义特异性的现象，提供免费校准信号。

👨‍🔬 Ziwei Su, Junyu Ren, Victor Veitch

🤖 AI资讯日报

📊 今日趋势总结

MIT Non-AI License

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

Why Boring Businesses Outlast AI Hype Cycles

The AI Crackpot Index

Ask HN: Anyone concerned about NYC Local Law 144?

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

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

NLP, AI, ML, bots – a passing trend or much more?

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

Show HN: Startup Raising capital through Book Sales

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

Bioinformatician

Pessimism's Paradox: Conservative Offline Training Amplifies Reward Hacking During Online Adaptation in Reasoning Models

Self-Evolving World Models for LLM Agent Planning

Words Speak Louder Than Code: Investigating Cognitive Heuristics in LLM-Based Code Vulnerability Detection

VLK: Learning Humanoid Loco-Manipulation from Synthetic Interactions in Reconstructed Scenes

C$^{2}$R: Cross-sample Consistency Regularization Mitigates Feature Splitting and Absorption in Sparse Autoencoders

MESA: Prioritizing Vulnerable Communication Channels for Securing Multi-Agent Systems

GROW$^2$: Grounding Which and Where for Robot Tool Use

DOPD: Dual On-policy Distillation

Beyond 2D Matching: A Unified Single-Stage Framework for Geometry-Aware Cross-View Object Geo-Localization

LeVo 2: Stable and Melodious Song Generation via Hierarchical Representation Modeling and Progressive Post-Training

A Multi-task Mixture of Experts Framework for Malware Classification, Packing Detection, and Family Attribution

Optimization Dynamics Imprint Semantic Specificity in Contrastive Embedding Norms

📅 历史日报目录