- 学术相关
- How recommendation affects customer search: a field experiment
- The life cycle of products: evidence and implications
- Are two heads better than one in AI-assisted decision making? Comparing the behavior and performance of groups and individuals in human-AI collaborative recidivism risk assessment
- 大模型驱动的自主智能体与群体智能
- A Survey on Large Language Model-Based Game Agents
- 复杂经济学
- 业界动态
- 技术技巧
学术相关
How recommendation affects customer search: a field experiment1
The life cycle of products: evidence and implications2
Are two heads better than one in AI-assisted decision making? Comparing the behavior and performance of groups and individuals in human-AI collaborative recidivism risk assessment3
作者没有社会科学研究经验，很多处理不严谨，结论仅供参考。
研究背景是罪犯再犯风险分析，使用的AI是COMPAS
COMPAS是基于机器学习分析大量罪犯犯罪记录、个人背景、社会关系、心理状况等信息分析再犯风险
COMPAS曾在美国被广泛使用，但现在被认为存在偏见、不透明、侵犯隐私等问题
- 与个体-AI协同决策相比，群体-AI协同决策中会更依赖AI给出的建议，无论AI是否正确
- 当个体/群体反对AI给出的建议时，群体比个体对自己的观点更自信
- 群体-AI协同决策比个体-AI协同决策更公平
- 当AI作出正确决定时，群体愿意给予AI更多信任
大模型驱动的自主智能体与群体智能
AgentVerse: 智能体协作平台4
A Survey on Large Language Model-Based Game Agents
- Simulation Games
- Human/Social Simulation
- Embodied Simulation
- Competition Games
- Cooperation Games
- Communication (Conversational) Games
复杂经济学
业界动态
人大陈旭丨Al Agent–大模型时代重要落地方向
在模拟过程中，效率是个非常重要的问题
技术技巧
Stanford 大语言模型赋能的知识策展系统——STORM
【Paper】5
它会针对一个主题基于互联网搜索，生成完整的类似于维基百科的长度报告，并附带引用。
工作流
- 写作前阶段：基于互联网搜索收集相关参考文献并生成大纲
- 写作阶段：基于大纲和参考文献生成带有引用的完整文章
针对文章问题的提示词自动优化：
- 观点引导的问题提问：给定输入主题，STORM通过调查类似主题的现有文章收集不同观点，并通过他们控制提问过程
- 模拟对话：STORM基于互联网资源，模拟维基百科作者与话题专家之间的对话，进而提升LLM对话题的理解能力并提出后续问题
-
Yuan, Z., Chen, A. Y., Wang, Y., & Sun, T. (2024). How recommendation affects customer search: A field experiment. Information Systems Research. https://doi.org/10.1287/isre.2022.0294 ↩
-
Argente, D., Lee, M., & Moreira, S. (2024). The life cycle of products: Evidence and implications. Journal of Political Economy, 132(2), 337–390. https://doi.org/10.1086/726704 ↩
-
Chiang, C.-W., Lu, Z., Li, Z., & Yin, M. (2023). Are two heads better than one in AI-assisted decision making? Comparing the behavior and performance of groups and individuals in human-AI collaborative recidivism risk assessment. Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, 1–18. https://doi.org/10.1145/3544548.3581015 ↩
-
Chen, W., Su, Y., Zuo, J., Yang, C., Yuan, C., Chan, C.-M., Yu, H., Lu, Y., Hung, Y.-H., Qian, C., Qin, Y., Cong, X., Xie, R., Liu, Z., Sun, M., & Zhou, J. (2023). AgentVerse: Facilitating multi-agent collaboration and exploring emergent behaviors (arXiv:2308.10848). arXiv. https://doi.org/10.48550/arXiv.2308.10848 ↩
-
Shao, Y., Jiang, Y., Kanell, T. A., Xu, P., Khattab, O., & Lam, M. S. (2024). Assisting in writing wikipedia-like articles from scratch with large language models (arXiv:2402.14207). arXiv. https://doi.org/10.48550/arXiv.2402.14207 ↩