{
  "date": "2026-09-26",
  "category": "Tech Daily",
  "papers": [
    {
      "id": "2602.19526",
      "title": "How to Train Your Deep Research Agent? Prompt, Reward, and Policy Optimization in Search-R1",
      "authors": ["Yinuo Xu", "Shuo Lu", "Jianjie Cheng", "Meng Wang", "Qianlong Xie", "Xingxing Wang", "Ran He", "Jian Liang"],
      "url": "https://arxiv.org/abs/2602.19526",
      "published": "2026-02-23",
      "summary": "Systematically studies RL for Deep Research agents across prompt template, reward function, and policy optimization. Demonstrates that Fast Thinking templates, action-level penalty mitigation for F1 rewards, and REINFORCE optimization enable Search-R1++ to substantially outperform prior baselines.",
      "apa_7": "Xu, Y., Lu, S., Cheng, J., Wang, M., Xie, Q., Wang, X., He, R., & Liang, J. (2026). How to train your deep research agent? Prompt, reward, and policy optimization in Search-R1. arXiv preprint arXiv:2602.19526. https://doi.org/10.48550/arXiv.2602.19526"
    },
    {
      "id": "2603.01050",
      "title": "MM-DeepResearch: A Simple and Effective Multimodal Agentic Search Baseline",
      "authors": ["Huanjin Yao", "Qixiang Yin", "Min Yang", "Ziwang Zhao", "Yibo Wang", "Haotian Luo", "Jingyi Zhang", "Jiaxing Huang"],
      "url": "https://arxiv.org/abs/2603.01050",
      "published": "2026-03-01",
      "summary": "Develops a multimodal deep research agent framework with explicit reasoning, multi-tool invocation, and cross-modal evidence synthesis, bridging the vision-language gap in autonomous research.",
      "apa_7": "Yao, H., Yin, Q., Yang, M., Zhao, Z., Wang, Y., Luo, H., Zhang, J., & Huang, J. (2026). MM-DeepResearch: A simple and effective multimodal agentic search baseline. arXiv preprint arXiv:2603.01050. https://doi.org/10.48550/arXiv.2603.01050"
    },
    {
      "id": "2602.05975",
      "title": "SAGE: Benchmarking and Improving Retrieval for Deep Research Agents",
      "authors": ["Tiansheng Hu", "Yilun Zhao", "Canyu Zhang", "Arman Cohan", "Chen Zhao"],
      "url": "https://arxiv.org/abs/2602.05975",
      "published": "2026-02-05",
      "summary": "Investigates and benchmarks LLM-based retrievers within multi-round deep research pipelines, demonstrating that iterative query reformulation and calibrated dense retrieval significantly boost task completion.",
      "apa_7": "Hu, T., Zhao, Y., Zhang, C., Cohan, A., & Zhao, C. (2026). SAGE: Benchmarking and improving retrieval for deep research agents. arXiv preprint arXiv:2602.05975. https://doi.org/10.48550/arXiv.2602.05975"
    },
    {
      "id": "2604.05952",
      "title": "Towards Trustworthy Report Generation: A Deep Research Agent with Progressive Confidence Estimation and Calibration",
      "authors": ["Yi Yuan", "Xuhong Wang", "Shanzhe Lei"],
      "url": "https://arxiv.org/abs/2604.05952",
      "published": "2026-04-07",
      "summary": "Proposes a deep research agent architecture integrating progressive confidence estimation and dynamic calibration to suppress hallucinations and overclaiming in automated research report generation.",
      "apa_7": "Yuan, Y., Wang, X., & Lei, S. (2026). Towards trustworthy report generation: A deep research agent with progressive confidence estimation and calibration. arXiv preprint arXiv:2604.05952. https://doi.org/10.48550/arXiv.2604.05952"
    },
    {
      "id": "2602.06052",
      "title": "A Survey of Agent Memory in the Second Half: Towards Self-Evolving and Long-Horizon Agents",
      "authors": ["Wei-Chieh Huang", "Weizhi Zhang", "Yueqing Liang", "Yuanchen Bei", "Yankai Chen", "Tao Feng", "Xinyu Pan", "Zhen Tan", "Yu Wang", "Tianxin Wei", "Kai Shu"],
      "url": "https://arxiv.org/abs/2602.06052",
      "published": "2026-01-14",
      "summary": "Provides a comprehensive taxonomy of agent memory systems in the second half of AI research, analyzing single and multi-agent memory topologies, trainable memory management policies, and test-time context curation.",
      "apa_7": "Huang, W. C., Zhang, W., Liang, Y., Bei, Y., Chen, Y., Feng, T., Pan, X., Tan, Z., Wang, Y., Wei, T., Wu, S., Xu, R., Yang, L., Yang, R., Yang, W., Yeh, C. Y., Zhang, H., Zhang, H., Zhu, S., ... Shu, K. (2026). A survey of agent memory in the second half: Towards self-evolving and long-horizon agents. arXiv preprint arXiv:2602.06052. https://doi.org/10.48550/arXiv.2602.06052"
    }
  ],
  "graft": {
    "repo": "https://github.com/stanford-oval/storm",
    "map_path": "profiles/r-and-d/raw_evidence/storm_map.txt",
    "saved_tokens": 139478,
    "parsed_files": 52,
    "symbols": 669,
    "edges": 1602
  }
}