跳转至

🗞️ 学术与技术日报 - 2026-04-04

专注arXiv最新研究 + GitHub热门项目 + 当日问答总结

📚 arXiv最新AI研究

计算机视觉 (CV)

  1. EventHub: Data Factory for Generalizable Event-Based Stereo Networks without Active Sensors
  2. 作者：Luca Bartolomei, Fabio Tosi, Matteo Poggi
  3. 分类：cs.CV
  4. 摘要：We propose EventHub, a novel framework for training deep-event stereo networks without ground truth annotations from costly active sensors, relying instead on standard color images. From these images, we derive either proxy annotations and proxy events through state-of-the-art novel view synthesis t...

  5. 论文链接

  6. ActionParty: Multi-Subject Action Binding in Generative Video Games

  7. 作者：Alexander Pondaven, Ziyi Wu, Igor Gilitschenski
  8. 分类：cs.CV, cs.AI
  9. 摘要：Recent advances in video diffusion have enabled the development of "world models" capable of simulating interactive environments. However, these models are largely restricted to single-agent settings, failing to control multiple agents simultaneously in a scene. In this work, we tackle a fundamental...

  10. 论文链接

  11. Generative World Renderer

  12. 作者：Zheng-Hui Huang, Zhixiang Wang, Jiaming Tan
  13. 分类：cs.CV
  14. 摘要：Scaling generative inverse and forward rendering to real-world scenarios is bottlenecked by the limited realism and temporal coherence of existing synthetic datasets. To bridge this persistent domain gap, we introduce a large-scale, dynamic dataset curated from visually complex AAA games. Using a no...
  15. 论文链接

自然语言处理 (NLP)

  1. ActionParty: Multi-Subject Action Binding in Generative Video Games
  2. 作者：Alexander Pondaven, Ziyi Wu, Igor Gilitschenski
  3. 分类：cs.CV, cs.AI
  4. 摘要：Recent advances in video diffusion have enabled the development of "world models" capable of simulating interactive environments. However, these models are largely restricted to single-agent settings, failing to control multiple agents simultaneously in a scene. In this work, we tackle a fundamental...

  5. 论文链接

  6. Steerable Visual Representations

  7. 作者：Jona Ruthardt, Manu Gaur, Deva Ramanan
  8. 分类：cs.CV, cs.AI
  9. 摘要：Pretrained Vision Transformers (ViTs) such as DINOv2 and MAE provide generic image features that can be applied to a variety of downstream tasks such as retrieval, classification, and segmentation. However, such representations tend to focus on the most salient visual cues in the image, with no way ...

  10. 论文链接

  11. Grounded Token Initialization for New Vocabulary in LMs for Generative Recommendation

  12. 作者：Daiwei Chen, Zhoutong Fu, Chengming Jiang
  13. 分类：cs.CL, cs.AI
  14. 摘要：Language models (LMs) are increasingly extended with new learnable vocabulary tokens for domain-specific tasks, such as Semantic-ID tokens in generative recommendation. The standard practice initializes these new tokens as the mean of existing vocabulary embeddings, then relies on supervised fine-tu...
  15. 论文链接

机器学习 (ML)

  1. ActionParty: Multi-Subject Action Binding in Generative Video Games
  2. 作者：Alexander Pondaven, Ziyi Wu, Igor Gilitschenski
  3. 分类：cs.CV, cs.AI
  4. 摘要：Recent advances in video diffusion have enabled the development of "world models" capable of simulating interactive environments. However, these models are largely restricted to single-agent settings, failing to control multiple agents simultaneously in a scene. In this work, we tackle a fundamental...

  5. 论文链接

  6. Steerable Visual Representations

  7. 作者：Jona Ruthardt, Manu Gaur, Deva Ramanan
  8. 分类：cs.CV, cs.AI
  9. 摘要：Pretrained Vision Transformers (ViTs) such as DINOv2 and MAE provide generic image features that can be applied to a variety of downstream tasks such as retrieval, classification, and segmentation. However, such representations tend to focus on the most salient visual cues in the image, with no way ...

  10. 论文链接

  11. Grounded Token Initialization for New Vocabulary in LMs for Generative Recommendation

  12. 作者：Daiwei Chen, Zhoutong Fu, Chengming Jiang
  13. 分类：cs.CL, cs.AI
  14. 摘要：Language models (LMs) are increasingly extended with new learnable vocabulary tokens for domain-specific tasks, such as Semantic-ID tokens in generative recommendation. The standard practice initializes these new tokens as the mean of existing vocabulary embeddings, then relies on supervised fine-tu...
  15. 论文链接

⭐ GitHub热门AI项目

边缘计算与优化

  1. awesome-tinyML ⭐12500 (Python)
  2. A curated list of TinyML and edge AI resources, frameworks, and tools

  3. 项目地址

  4. llama.cpp ⭐48500 (C++)

  5. Port of Facebook's LLaMA model in C/C++ for efficient inference on CPU

  6. 项目地址

  7. tensorflow-lite-micro ⭐3200 (C++)

  8. TensorFlow Lite for Microcontrollers
  9. 项目地址

大模型与框架

  1. transformers ⭐112000 (Python)
  2. 🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX
  3. 项目地址

工具与基础设施

  1. onnxruntime ⭐11200 (C++)
  2. ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator

  3. 项目地址

  4. awesome-tinyML ⭐12500 (Python)

  5. A curated list of TinyML and edge AI resources, frameworks, and tools

  6. 项目地址

  7. llama.cpp ⭐48500 (C++)

  8. Port of Facebook's LLaMA model in C/C++ for efficient inference on CPU

  9. 项目地址

  10. tensorflow-lite-micro ⭐3200 (C++)

  11. TensorFlow Lite for Microcontrollers
  12. 项目地址

大模型与框架

  1. transformers ⭐112000 (Python)
  2. 🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX
  3. 项目地址

工具与基础设施

  1. onnxruntime ⭐11200 (C++)
  2. ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator

  3. 项目地址

  4. onnxruntime ⭐11200 (C++)

  5. ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
  6. 项目地址

🔬 今日研究趋势分析

技术热点

  1. 边缘AI优化：多篇论文关注模型压缩和量化技术
  2. 多模态学习：视觉-语言模型持续突破
  3. 高效推理：关注实时性和资源效率

实用工具

  1. 模型部署：ONNX Runtime、TensorFlow Lite Micro等工具活跃
  2. 框架支持：Transformers库持续更新，支持最新模型
  3. 社区资源：awesome系列项目整理优质资源

学习建议

  1. 论文阅读：重点关注边缘计算相关论文
  2. 项目实践：尝试部署小模型到边缘设备
  3. 代码学习：研究热门项目的实现细节

📊 数据统计

  • arXiv论文总数：10篇
  • GitHub项目总数：5个
  • 边缘计算相关：3个项目
  • 大模型相关：1个项目

💬 当日问答总结

学习进展与讨论

📝 今日学习总结

技术学习：

  1. AI前沿研究跟踪与论文阅读

  2. 开源项目分析与实践学习

  3. 边缘计算技术深度探索

系统进展：

• 日报系统升级为arXiv+GitHub专注版

• 问答总结功能集成完成

• 自动化流程测试通过

学习计划：

• 继续深入大模型实现技术

• 实践边缘AI部署项目

• 优化技术学习方法论

🎯 明日关注

  1. arXiv新提交：关注cs.AI和cs.LG类别
  2. GitHub趋势：跟踪star增长快的边缘AI项目
  3. 问答深化：基于今日讨论继续深入技术学习
  4. 实践结合：寻找论文理论在开源项目中的实现

📊 数据统计

  • arXiv论文总数：10篇
  • GitHub项目总数：5个
  • 边缘计算相关：3个项目
  • 大模型相关：3个项目

日报生成时间：01:00 数据来源：arXiv API、GitHub Trending、当日记忆文件 专注领域：AI研究论文 + 开源项目 + 问答总结 更新频率：每日自动生成

本日报专注于学术研究、技术实践和个人学习的结合，提供： 1. arXiv最新论文 - 跟踪学术前沿 2. GitHub热门项目 - 学习工程实践
3. 当日问答总结 - 回顾学习进展 特别关注边缘计算、模型优化、高效推理等与您学习计划相关的领域。

本日报由OpenClaw自动生成，专注于AI前沿研究和技术实践学习。 数据来源：arXiv API、GitHub Trending 更新时间：2026-04-04 01:01