博文

TST征稿｜推进智能体大模型：基础、能力与新兴应用

已有 177 次阅读 2026-7-28 09:18 |个人分类:学报英文版|系统分类:论文交流

The rapid evolution of large-scale AI models has led to a paradigm shift from passive, instruction-following systems toward Agentic Large Models (ALMs) that possess autonomous reasoning, interactive decision‑making, strategic planning, and adaptive problem‑solving capabilities. As ALMs increasingly integrate perception, cognition, memory, and action into unified intelligent systems, unlocking their full potential has become a central challenge in AI research.

Despite remarkable progress, building effective and reliable agentic models requires significant advances in behavioral alignment, multi-step reasoning, environmental adaptation, tool-use proficiency, multi-agent collaboration, and safety assurance. At the same time, real-world deployment across science, industry, robotics, communications, and general-purpose automation demands new approaches to enhance capability generalization, operational robustness, and domain transferability.

This Special Issue aims to explore new learning paradigms, system architectures, and practical breakthroughs that push the boundaries of agentic large models. We welcome original research that deepens our understanding of ALMs’ cognitive mechanisms, develops new methods to enhance their agency, or demonstrates innovative applications that reveal emerging opportunities and challenges.

1. Scope of Topics

Topics of interest include, but are not limited to:

Foundations and architectures for agentic large models

Cognitive mechanisms: planning, memory, tool use, and long‑horizon reasoning

Adaptive and interactive behaviors in dynamic or uncertain environments

Multi-agent coordination, communication, and emergent collaboration

Behavioral alignment, safety, interpretability, and trustworthiness in ALMs

Data and knowledge integration for enhancing agency and task performance

Evaluation frameworks and benchmarks for autonomous intelligence

ALMs for robotics, embodied agents, and real-world decision systems

ALMs for scientific discovery, engineering optimization, and complex simulations

ALMs in communication systems, automation, and cyber–physical environments

Resource-efficient architectures for scalable agentic intelligence

Applications and case studies showcasing breakthroughs using agentic models

2. Submission Guidelines

Authors should prepare papers in accordance with the format requirements of Tsinghua Science and Technology, with reference to the Instruction given at https://www.sciopen.com/journal/1007-0214, and submit the complete manuscript through the online manuscript submission system at https://mc03.manuscriptcentral.com/tst with manuscript type as “Special Issue on Advancing Agentic Large Models: Foundations, Capabilities, and Emerging Applications”.

3. Important Dates

Deadline for submissions: May 31, 2026

4. Guest Editors

Jing Zhang, Renmin University of China

Xu Chen，Renmin University of China

Linmei Hu，Beijing Institute of Technology

Tsinghua Science and Technology 期刊介绍

Tsinghua Science and Technology是清华大学主办的第一本自然科学类英文期刊。2011年，本刊改版为信息科学类专业期刊，主要瞄准国家创新发展关键领域和战略方向，立足信息科学领域全球最新研究成果，全面反映人工智能、大数据、信息与通信工程、控制科学与工程、计算机科学与技术、软件工程等方面最新原创性研究成果，旨在为信息科学的研究和发展搭建了国际化学术交流平台。专业化转型后，本刊进入快速发展阶段，在学术质量及国际影响力上都得到了很大的提升，陆续被SCIE、SCOPUS、EI、CSCD等国内外数据库收录；2018年，期刊荣获“第四届中国出版政府奖期刊提名奖”；2019~2023年入选“中国科技期刊卓越行动计划”梯队期刊项目；2024年入选中国科技期刊卓越行动计划二期英文领军期刊项目；2025年入选北京市“2025支持高水平国际科技期刊建设-强刊提升类”项目。

经过十余年的努力，本刊受到了全球信息科学领域学者认可，国际稿源稳步提升，读者来自全球130多个国家及地区。2025年8月，Tsinghua Science and Technology第31卷第1期正式在清华大学出版社SciOpen平台出版发布，标志着Tsinghua Science and Technology结束了“借船出海”的办刊模式，回归本土独立运营！

SciOpen是由清华大学出版社自主研发的科技期刊国际化数字出版平台，于2022年6月正式上线。截至目前，SciOpen平台目前已汇聚195种期刊，上线超过5.8万篇文章，2025年7月访问量突破960万次，用户遍布190多个国家和地区。SciOpen平台已实现与国际通用搜索引擎、出版标准化组织及重要学术索引数据库的全面对接，运行稳定、服务可靠，为Tsinghua Science and Technology回归后的持续发展提供了坚实技术支撑和全球化传播保障。

未来，Tsinghua Science and Technology将继续聚焦信息科学前沿和热点研究，依托SciOpen平台先进的服务能力，为我国乃至全球的信息科学领域发展贡献更多的智慧与力量。