AI Agent

Patience is all you need! An agentic system for performing scientific literature review

David Brett, Anniek Myatt

专题命中 工具调用 ：agentic(title)；分类 cs.AI

Comments 10 pages, 5 figures

Agentic AI for Improving Precision in Identifying Contributions to Sustainable Development Goals

William A. Ingram, Bipasha Banerjee, Edward A. Fox

专题命中 工具调用 ：agentic(title)；分类 cs.AI

Recurrent Auto-Encoders for Enhanced Deep Reinforcement Learning in Wilderness Search and Rescue Planning

Jan-Hendrik Ewers, David Anderson, Douglas Thomson

专题命中 工具调用 ：planning(title)；分类 cs.LG

Comments Submitted to Machine Learning with Applications

SibylSat: Using SAT as an Oracle to Perform a Greedy Search on TOHTN Planning

Gaspard Quenard, Damier Pellier, Humbert Fiorino

专题命中 工具调用 ：planning(title)；分类 cs.AI

Journal ref ECAI 2024, Oct 2024, Santiago de Compostela, Spain. pp.4157 - 4164

A Study on the Implementation Method of an Agent-Based Advanced RAG System Using Graph

Cheonsu Jeong

专题命中 工具调用 ：agent(title)；分类 cs.AI

Journal ref 2024 Knowledge Management Research

Jack of All Trades, Master of Some, a Multi-Purpose Transformer Agent

Quentin Gallouédec, Edward Beeching, Clément Romac, Emmanuel Dellandréa

专题命中 工具调用 ：agent(title)；分类 cs.AI

Journal ref 38th Workshop on Aligning Reinforcement Learning Experimentalists and Theorists (ARLET 2024)

AI Tool Use and Adoption in Software Development by Individuals and Organizations: A Grounded Theory Study

Ze Shi Li, Nowshin Nawar Arony, Ahmed Musa Awon, Daniela Damian, Bowen Xu

专题命中 工具调用 ：tool use(title)；分类 cs.SE

Budget-Constrained Tool Learning with Planning

Yuanhang Zheng, Peng Li, Ming Yan, Ji Zhang, Fei Huang, Yang Liu

专题命中 工具调用 ：planning(title)；分类 cs.AI

Comments Accepted for Findings of ACL 2024

Modeling Collaborator: Enabling Subjective Vision Classification With Minimal Human Effort via LLM Tool-Use

Imad Eddine Toubal, Aditya Avinash, Neil Gordon Alldrin, Jan Dlabal, Wenlei Zhou, Enming Luo, Otilia Stretcu, Hao Xiong, Chun-Ta Lu, Howard Zhou, Ranjay Krishna, Ariel Fuxman, Tom Duerig

专题命中 工具调用 ：tool-use(title)；分类 cs.LG

Equipping Language Models with Tool Use Capability for Tabular Data Analysis in Finance

Adrian Theuma, Ehsan Shareghi

专题命中 工具调用 ：tool use(title)；分类 cs.CL

Comments Accepted to EACL2024; code, model and dataset are available at https://raven-lm.github.io

Online search of unknown terrains using a dynamical system-based path planning approach

Karan Sridharan, Patrick McNamee, Zahra Nili Ahmadabadi, Jeffrey Hudack

专题命中 工具调用 ：planning(title)；分类 cs.AI

Journal ref J Intell Robot Syst 106, 21 (2022)

A Deep Neural Networks ensemble workflow from hyperparameter search to inference leveraging GPU clusters

Pierrick Pochelu, Serge G. Petiton, Bruno Conche

专题命中 工具调用 ：workflow(title)；分类 cs.LG

Journal ref ACM International Conference Proceeding Series 2022

Designing an Automatic Agent for Repeated Language based Persuasion Games

Maya Raifer, Guy Rotman, Reut Apel, Moshe Tennenholtz, Roi Reichart

专题命中 工具调用 ：agent(title)；分类 cs.CL

Comments Accepted for TACL in December 2021

Improving Confidence in the Estimation of Values and Norms

Luciano Cavalcante Siebert, Rijk Mercuur, Virginia Dignum, Jeroen van den Hoven, Catholijn Jonker

专题命中 工具调用 ：agent(abstract,comments)；autonomous agent(abstract)；分类 cs.AI；multi-agent(comments)

Comments 16 pages, 3 figures, pre-print for the International Workshop on Coordination, Organizations, Institutions, Norms and Ethics for Governance of Multi-Agent Systems (COINE), co-located with AAMAS 2020

MEEP: An Open-Source Platform for Human-Human Dialog Collection and End-to-End Agent Training

Arkady Arkhangorodsky, Amittai Axelrod, Christopher Chu, Scot Fang, Yiqi Huang, Ajay Nagesh, Xing Shi, Boliang Zhang, Kevin Knight

专题命中 工具调用 ：agent(title)；分类 cs.CL

Comments 10 pages

Strategic API Analysis and Planning: APIS Technical Report

Jennifer Horkoff, Juho Lindman, Imed Hammouda, Eric Knauss

专题命中 工具调用 ：planning(title)；分类 cs.SE

Improved Learning in Evolution Strategies via Sparser Inter-Agent Network Topologies

Dhaval Adjodah, Dan Calacci, Yan Leng, Peter Krafft, Esteban Moro, Alex Pentland

专题命中 工具调用 ：agent(title)；分类 cs.AI

Comments This paper is obsolete

The N-Tuple Bandit Evolutionary Algorithm for Game Agent Optimisation

Simon M Lucas, Jialin Liu, Diego Perez-Liebana

专题命中 工具调用 ：agent(title)；分类 cs.AI

Comments 9 pages, 3 figures, 3 table. This is the final version of the article accepted by WCCI2018

Towards Real-Time Search Planning in Subsea Environments

James McMahon, Harun Yetkin, Artur Wolek, Zachary Waters, Dan Stilwell

专题命中 工具调用 ：planning(title)；分类 cs.AI

Comments 8 pages, 5 figures. Submitted to 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2017)

Path planning with Inventory-driven Jump-Point-Search

Davide Aversa, Sebastian Sardina, Stavros Vassos

专题命中 工具调用 ：planning(title)；分类 cs.AI

Journal ref In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE), pp. 2-8, 2015

Neurohex: A Deep Q-learning Hex Agent

Kenny Young, Ryan Hayward, Gautham Vasan

专题命中 工具调用 ：agent(title)；分类 cs.AI

An Agent based Approach towards Metadata Extraction, Modelling and Information Retrieval over the Web

Zeeshan Ahmed, Detlef Gerhard

专题命中 工具调用 ：agent(title)；分类 cs.AI

Comments In the proceedings of First International Workshop on Cultural Heritage on the Semantic Web in conjunction with the 6th International Semantic Web Conference and the 2nd Asian Semantic Web Conference 2007, (ISWC + ASWC 2007), P 117, 12-15 November 2007

PeakBench: Benchmarking Resource-Aware Tool Invocation in LLM Agents

PeakBench：面向大语言模型智能体的资源感知工具调用基准测试

Zhi-Kai Chen, Xu-Xiang Zhong, Song-Yan Li, De-Chuan Zhan, Han-Jia Ye

机构 * School of Artificial Intelligence, Nanjing University（南京大学人工智能学院） ； Nanjing University（南京大学） 纠错

专题命中 工具调用 ：agent(abstract)；planning(abstract)；分类 cs.AI、cs.SE

AI总结 PeakBench是面向LLM智能体的资源感知工具调用基准，通过两部分评估框架解耦逻辑规划与物理调度，实验表明资源信息可减少溢出、提升利用率，为相关研究提供测试平台。

ToolRobustBench: Stage-Wise Perturbation Evaluation and Failure Diagnosis for Tool-Calling Agents

ToolRobustBench：工具调用智能体的分阶段扰动评估与故障诊断

YiShan Zheng, Yuan Wu, Yi Chang

专题命中 工具调用 ：agent(abstract)；tool-use(abstract)；分类 cs.AI、cs.SE

AI总结 本研究推出ToolRobustBench，这一工具调用智能体的分阶段诊断基准，通过四类扰动族评估7个模型等的15456个实例，发现工具输出/观测扰动是主要瓶颈，可诊断工具调用故障来源与传播。

Enrich-Retrieve-Rank: Scaling Capability Discovery Beyond In-Context Routing

Enrich-Retrieve-Rank：将能力发现扩展至上下文路由之外

Nazib Sorathiya, Daniel Zhang, Bardiya Akhbari

机构 * Amazon AGI（亚马逊AGI） 纠错

专题命中 工具调用 ：agent(abstract)；multi-agent(abstract)；分类 cs.AI、cs.CL

AI总结 该研究提出Enrich-Retrieve-Rank流程，将能力发现从上下文路由扩展，经实验验证其在大规模MATS组件场景下，比Full-Ctx、Search&Pick基线性能更优且成本更低，已作为多智能体平台的默认能力发现层投入生产。

Comments 11 pages, 4 figures, and 12 tables

Chain of Operators: An Inference-Time Harness for In-Context Operator Learning

利用算子链实现上下文算子学习

Minghui Yang, Chenghan Wu, Ling Guo, Liu Yang

机构 * Department of Mathematics, Shanghai Normal University（上海师范大学数学系） ； Department of Mathematics, National University of Singapore（新加坡国立大学数学系） 纠错

专题命中 工具调用 ：tool use(abstract)；agentic(abstract)；分类 cs.AI、cs.LG

AI总结 提出Chain of Operators (CHOP)框架，通过构造显式初等变换与冻结ICON的算子链，无需微调即可提升上下文算子网络在分布外算子任务上的泛化能力，在标量守恒律和平均场控制问题中降低推理误差。

Same Payload, Different Channel: Measuring Trust Asymmetry in Tool-Using Language Models

相同载荷，不同通道：测量使用工具的語言模型中的信任不对称性

Mohammed Sameer Syed, Rozhin Yasaei

机构 * University of Arizona（亚利桑那大学） 纠错

专题命中 工具调用 ：agent(abstract)；agentic(abstract)；分类 cs.CL、cs.LG

AI总结 本研究提出安全不对称分数（SAS），通过匹配恶意载荷仅改变传递上下文，系统测量了语言模型在不同通道（用户消息、工具元数据、工具输出）中对对抗性内容的脆弱性差异，发现代理原生模型在工具描述通道更脆弱，而通用模型相反，且机制研究表明安全相关表示在深层网络非线性编码。

Comments Accepted to EMNLP 2026 (Main Conference)

Safety Training May Persist Through Helpfulness Optimization in LLM Agents

在LLM代理中通过帮助性优化保持安全性训练

Benjamin Plaut

机构 * Department of Computer Science, University of California, Berkeley, USA（计算机科学系，加州大学伯克利分校） 纠错

专题命中 工具调用 ：tool-use(abstract)；agentic(abstract)；分类 cs.CL、cs.LG

AI总结 研究通过帮助性优化在LLM代理中保持安全性训练，发现单独训练或依次训练无法找到最佳策略，但所有配置接近帕累托前沿。

Comments Preprint

IACDM: Interactive Adversarial Convergence Development Methodology -- A Structured Framework for AI-Assisted Software Development

IACDM：交互对抗收敛开发方法论——面向AI辅助软件开发的结构化框架

Jasmine Moreira

专题命中 工具调用 ：agent(abstract)；tool use(abstract)；分类 cs.AI、cs.SE

AI总结 本文提出IACDM方法论，通过外部验证代理解决AI生成应用中的验证缺口问题，强调通过层次语义分析、知识管理及系统对抗批评提升软件开发质量。

Comments 37 pages, 7 tables. Technical Foundation Document. v3 adds a pre-registered experiment testing lens non-redundancy over 12 projects, and withdraws the retrospective analysis of v1-v2 (defective instrument). Data: https://doi.org/10.5281/zenodo.21908908 Repo: https://github.com/jasminemoreira/Versus

Automated Design Optimization via Strategic Search with Large Language Models

通过大语言模型的战略搜索实现自动化设计优化

Anthony Carreon, Vansh Sharma, Venkat Raman

专题命中 工具调用 ：agent(abstract)；planning(abstract)；分类 cs.AI、cs.LG

AI总结 本研究提出AUTO框架，利用大语言模型的战略搜索实现GPU代码优化，提升搜索效率并降低成本。

Comments 16 pages, 4 tables, 8 figures, preprint