Papers

We introduce the notion of width-sketching algorithms, defined as algorithms with provably bounded width (that is, probability of containing the randomness) for the induced coverage set. For algorithms that sketch the width, we prove a novel uniform upper bound and provide an instance where the width in expectation is twice as large as the optimal width. We then introduce the width-optimality notion and an approximate version termed mean-width optimality, which allows us to derive algorithms with the desired coverage while minimizing the mean width. We provide a high-level perspective on the relationship with depth-sketching algorithms, i.e., algorithms that sketch the depth of the induced sets with probability 1 − α, and show that they provide complementary forms of coverage. Finally, we demonstrate the application of the framework to conformal prediction with Bayesian quadrature.

随着生成式人工智能与智能体（Agent）技术的迅猛发展，新闻传播领域正经历从"内容数字化" 向"认知智能化"的范式转型。受众智能体作为能够模拟、预测甚至替代部分人类受众认知与行为的新型数 字实体，其在新闻生产、分发与反馈各环节的深度嵌入，在提升传播效率的同时也引发了复杂的伦理挑战。 本文结合2025年斯坦福大学AI行为研究、中国AI大模型测评报告等最新实证数据，系统审视受众智能体 在新闻传播中的应用所衍生的伦理风险，并构建相应的规制路径。研究发现，受众智能体的伦理风险主要 集中在三个层面：在模拟层面，存在"数字孪生"失真、归因悖论与信任赤字的风险；在影响层面，面临商 业价值侵蚀公共属性、人机协同失当导致价值偏移的困境；在驯化层面，则遭遇技术依赖导致的主体性消 解与规则滞后带来的治理真空。针对上述风险，本文借鉴动态能力理论，提出一个以"感知-捕捉-重构"为 核心的多维治理框架，为新型主流媒体在智能时代的稳健变革提供兼具学理与实践价值的方案。

Definitions of the human span biology, psychology, anthropology, law, and philosophy, resisting reduction to a single trait. This study introduces HUMN-DEF, a multiaxial framework that models seven definitional axes—Taxonomic/Evolutionary (A1), Genetic/Developmental (A2), Cognitive/Linguistic (A3), Physiological/Regulatory (A4), Sociocultural/Anthropological (A5), Legal/Normative (A6), and Phenomenological/Subjective (A7)—and represents texts as Definition Profile Vectors (DPVs). A purposive cross-disciplinary corpus (n = 31) was coded by two independent automated procedures (Krippendorff’s α = .84), analyzed with post-stratification weights (field × decade × language), and evaluated via percentile bootstraps. Results converge on Sociocultural (A5) and Cognitive/Linguistic (A3) as predominant emphases; Taxonomy/Genetics (A1/A2) anchor but are not sufficient; Legal/Normative (A6) rises under balanced representation; Phenomenology (A7) is mid-level; Physiology (A4) is specialized. Cross-field disagreement, measured with a Definitional Diversity Index (Jensen–Shannon divergence), is moderate (0.394; 95% CIs ≈ [0.345, 0.475]). We argue that “human” is best treated as a transparent, context-weighted mixture over A1–A7.

This paper proposes a federated model of AI Scientists, integrating a layered stack architecture, an iterative discovery pipeline, and a governance-aligned roadmap. We argue that AI Scientists should not only accelerate discovery but also serve as custodians of epistemic integrity. Through case studies in drug discovery, climate modeling, and materials science, we demonstrate how federation enables cross-domain synthesis while embedding reproducibility, incentive alignment, and participatory governance. We conclude with a research roadmap toward Trusted AI Scientists, highlighting technical, incentive, and governance challenges.