High-fidelity agent initialization is crucial for credible Agent-Based Modeling across diverse domains.A robust framework should be Topic-Adaptive, capturing macro-level joint distributions while ensuring micro-level individual rationality.Existing approaches fall into two categories: static data-based retrieval methods that fail to adapt to unseen topics absent from the data, and LLM-based generation methods that lack macro-level distribution awareness, resulting in inconsistencies between micro-level persona attributes and reality.To address these problems, we propose HAG, a Hierarchical Agent Generation framework that formalizes population generation as a two-stage decision process.Firstly, utilizing a World Knowledge Model to infer hierarchical conditional probabilities to construct the Topic-Adaptive Tree, achieving macro-level distribution alignment.Then, grounded real-world data, instantiation and agentic augmentation are carried out to ensure micro-level consistency.Given the lack of specialized evaluation, we establish a multi-domain benchmark and a comprehensive PACE evaluation framework.Extensive experiments show that HAG significantly outperforms representative baselines, reducing population alignment errors by an average of 37.7% and enhancing sociological consistency by 18.8%.

Xiucheng Xu, Bingbing Xu, Tian Xueyun, Zihe Huang, Rongxin Chen, Li Yunfan, Huawei Shen. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Multi-personality generation for LLMs, enabling simultaneous embodiment of multiple personalization attributes, is a fundamental challenge. Existing retraining-based approaches are costly and poorly scalable, while decoding-time methods often rely on external models or heuristics, limiting flexibility and robustness. In this paper, we propose a novel Multi-Personality Generation (MPG) framework under the decoding-time combination paradigm. It flexibly controls multi-personality without relying on scarce multi-dimensional models or extra training, leveraging implicit density ratios in single-dimensional models as a ''free lunch'' to reformulate the task as sampling from a target strategy aggregating these ratios. To implement MPG efficiently, we design Speculative Chunk-level based Rejection sampling (SCR), which generates responses in chunks and parallelly validates them via estimated thresholds within a sliding window. This significantly reduces computational overhead while maintaining high-quality generation. Experiments on MBTI personality and Role-Playing demonstrate the effectiveness of MPG, showing improvements up to 16%–18%. Code and data are available at https://github.com/Libra117/MPG.

As artificial intelligence rapidly advances, addressing the interplay of technical, ethical, and risk factors in optimizing digital market decision-making through AI platforms has become increasingly prominent. However, the impact of these factors on market performance, particularly in investment value, remains underexplored. The study, based on 412 validated responses from service industry professionals gathered through a carefully designed questionnaire, aims to predict the relationship among these factors and their influence on market performance. It also explores how cognitive engagement mediates the relationship between AI platforms and financial metrics. Key findings:(1) the interplay of technical, ethical, and risk factors optimizes market decision-making and guides AI investments; (2) cognitive engagement, especially in the services sector, is essential to maximize the impact of AI platforms on market performance. The study provides valuable insights into AI's role in shaping market dynamics within the services sector and relevant governance recommendations for policymakers.

In the era of digitalization and green development, collaborative green innovation is gaining increasing attention. However, how digital transformation affects green innovation, especially collaboration in green innovation is understudied. Based on the data analysis of Chinese listed firms between 2008 and 2021, this study investigates the relationship between digital transformation and collaborative green innovation and explores further the contingencies of financial slack and market competition. Our findings are as follows: 1) digital transformation increases the likelihood of having collaborative green innovation; 2) financial slack enhances this positive relationship, while the moderating role of market competition is not significant. Overall, our study advances both green innovation and digital transformation literature by building a linkage between digital transformation, collaboration innovation, and green innovation.