AI资讯日报 - 2025/9/22

👨‍🔬 Jane Luo, Xin Zhang, Steven Liu, Jie Wu, Yiming Huang, Yangyu Huang, Chengyu Yin, Ying Xin, Jianfeng Liu, Yuefeng Zhan, Hao Sun, Qi Chen, Scarlett Li, Mao Yang

Beyond Pointwise Scores: Decomposed Criteria-Based Evaluation of LLM Responses

学术论文 ArXiv 重要度: 8

提出DeCE分解评估框架，将答案质量分解为精确度和召回率，在专业领域评估中与专家判断相关性达0.78。

👨‍🔬 Fangyi Yu, Nabeel Seedat, Dasha Herrmannova, Frank Schilder, Jonathan Richard Schwarz

DiffusionNFT: Online Diffusion Reinforcement with Forward Process

学术论文 ArXiv 重要度: 8

提出DiffusionNFT在线强化学习范式，通过前向过程优化扩散模型，效率比FlowGRPO提升25倍。

👨‍🔬 Kaiwen Zheng, Huayu Chen, Haotian Ye, Haoxiang Wang, Qinsheng Zhang, Kai Jiang, Hang Su, Stefano Ermon, Jun Zhu, Ming-Yu Liu

CultureScope: A Dimensional Lens for Probing Cultural Understanding in LLMs

学术论文 ArXiv 重要度: 7

提出CultureScope评估框架，基于文化冰山理论构建140维分类模式，系统评估LLM的文化理解能力。

👨‍🔬 Jinghao Zhang, Sihang Jiang, Shiwei Guo, Shisong Chen, Yanghua Xiao, Hongwei Feng, Jiaqing Liang, Minggui HE, Shimin Tao, Hongxia Ma

Robust Vision-Language Models via Tensor Decomposition: A Defense Against Adversarial Attacks

学术论文 ArXiv 重要度: 7

提出基于张量分解的轻量级防御方法，无需重新训练即可提升视觉语言模型对抗攻击的鲁棒性。

👨‍🔬 Het Patel, Muzammil Allie, Qian Zhang, Jia Chen, Evangelos E. Papalexakis

Attention Schema-based Attention Control (ASAC): A Cognitive-Inspired Approach for Attention Management in Transformers

学术论文 ArXiv 重要度: 7

基于注意力图式理论提出ASAC模块，通过VQVAE实现注意力精确管理，提升Transformer效率与鲁棒性。

👨‍🔬 Krati Saxena, Federico Jurado Ruiz, Guido Manzi, Dianbo Liu, Alex Lamb

See&Trek: Training-Free Spatial Prompting for Multimodal Large Language Model

学术论文 ArXiv 重要度: 6

提出SEE&TREK训练免费提示框架，通过视觉多样性和运动重建增强MLLM的空间理解能力。

👨‍🔬 Pengteng Li, Pinhao Song, Wuyang Li, Weiyu Guo, Huizai Yao, Yijie Xu, Dugang Liu, Hui Xiong

FocalCodec-Stream: Streaming Low-Bitrate Speech Coding via Causal Distillation

学术论文 ArXiv 重要度: 6

提出FocalCodec-Stream流式语音编解码器，在0.55-0.80kbps低比特率下实现高质量语音压缩。

👨‍🔬 Luca Della Libera, Cem Subakan, Mirco Ravanelli

Communications to Circulations: 3D Wind Field Retrieval and Real-Time Prediction Using 5G GNSS Signals and Deep Learning

学术论文 ArXiv 重要度: 6

提出G-WindCast框架，利用5G GNSS信号和深度学习实现三维风场检索与实时预测。

👨‍🔬 Yuchen Ye, Hong Liang, Chaoxia Yuan, Mingyu Li, Aoqi Zhou, Chunqing Shang, Hua Cai, Peixi Liu, Kezuan Wang, Yifeng Zheng

Network-Based Detection of Autism Spectrum Disorder Using Sustainable and Non-invasive Salivary Biomarkers

学术论文 ArXiv 重要度: 6

开发GANet网络优化框架，基于唾液光谱数据实现自闭症谱系障碍的非侵入性检测，准确率达78%。

👨‍🔬 Janayna M. Fernandes, Robinson Sabino-Silva, Murillo G. Carneiro

Accelerating Atomic Fine Structure Determination with Graph Reinforcement Learning

学术论文 ArXiv 重要度: 5

提出图强化学习方法自动化原子精细结构确定，将数年分析任务缩短至数小时完成。

👨‍🔬 M. Ding, V. -A. Darvariu, A. N. Ryabtsev, N. Hawes, J. C. Pickering

Fast OTSU Thresholding Using Bisection Method

学术论文 ArXiv 重要度: 4

利用二分法优化Otsu阈值算法，计算复杂度从O(L)降至O(log L)，保持分割精度。

👨‍🔬 Sai Varun Kodathala

🤖 AI资讯日报

📊 今日趋势总结

Why Boring Businesses Outlast AI Hype Cycles

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

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

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

Ask HN: Anyone concerned about NYC Local Law 144?

The AI Crackpot Index

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

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

Ask HN: Dipping my toes with artificial intelligence and what to expect? (CS)

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

Show HN: Startup Raising capital through Book Sales

Bioinformatician

Alibaba-NLP/DeepResearch

RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation

Beyond Pointwise Scores: Decomposed Criteria-Based Evaluation of LLM Responses

DiffusionNFT: Online Diffusion Reinforcement with Forward Process

CultureScope: A Dimensional Lens for Probing Cultural Understanding in LLMs

Robust Vision-Language Models via Tensor Decomposition: A Defense Against Adversarial Attacks

Attention Schema-based Attention Control (ASAC): A Cognitive-Inspired Approach for Attention Management in Transformers

See&Trek: Training-Free Spatial Prompting for Multimodal Large Language Model

FocalCodec-Stream: Streaming Low-Bitrate Speech Coding via Causal Distillation

Communications to Circulations: 3D Wind Field Retrieval and Real-Time Prediction Using 5G GNSS Signals and Deep Learning

Network-Based Detection of Autism Spectrum Disorder Using Sustainable and Non-invasive Salivary Biomarkers

Accelerating Atomic Fine Structure Determination with Graph Reinforcement Learning

Fast OTSU Thresholding Using Bisection Method

📅 历史日报目录