大模型推理能力

Characterizing Technical Debt and Antipatterns in AI-Based Systems: A Systematic Mapping Study

Justus Bogner, Roberto Verdecchia, Ilias Gerostathopoulos

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI、cs.LG

Comments Accepted at the 4th International Conference on Technical Debt (TechDebt 2021)

ERICA: Improving Entity and Relation Understanding for Pre-trained Language Models via Contrastive Learning

Yujia Qin, Yankai Lin, Ryuichi Takanobu, Zhiyuan Liu, Peng Li, Heng Ji, Minlie Huang, Maosong Sun, Jie Zhou

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Comments Accepted by ACL-IJCNLP 2021 main conference

Giving Commands to a Self-driving Car: A Multimodal Reasoner for Visual Grounding

Thierry Deruyttere, Guillem Collell, Marie-Francine Moens

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Comments Updated acknowledgements

Agent-Centric Representations for Multi-Agent Reinforcement Learning

Wenling Shang, Lasse Espeholt, Anton Raichuk, Tim Salimans

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI、cs.LG

Influencing Reinforcement Learning through Natural Language Guidance

Tasmia Tasrin, Md Sultan Al Nahian, Habarakadage Perera, Brent Harrison

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI、cs.LG

Comments 7 pages, 6 figures, The 34th International FLAIRS Conference, 2021

A Data-Driven Study of Commonsense Knowledge using the ConceptNet Knowledge Base

Ke Shen, Mayank Kejriwal

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Critical Thinking for Language Models

Gregor Betz, Christian Voigt, Kyle Richardson

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Empowering Things with Intelligence: A Survey of the Progress, Challenges, and Opportunities in Artificial Intelligence of Things

Jing Zhang, Dacheng Tao

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI、cs.LG

Comments Accepted by IEEE Internet of Things Journal

Stronger Transformers for Neural Multi-Hop Question Generation

Devendra Singh Sachan, Lingfei Wu, Mrinmaya Sachan, William Hamilton

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.LG

Comments Code will be made available

Challenge Closed-book Science Exam: A Meta-learning Based Question Answering System

Xinyue Zheng, Peng Wang, Qigang Wang, Zhongchao Shi

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Moral Dilemmas for Artificial Intelligence: a position paper on an application of Compositional Quantum Cognition

Camilo M. Signorelli, Xerxes D. Arsiwalla

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Comments 15 pages, 3 figures, Conference paper at Quantum Interaction 2018, Nice, France. Published in Lecture Notes in Computer Science, vol 11690, Springer, Cham. Online ISBN 978-3-030-35895-2

Journal ref Quantum Interaction. QI 2018. Lecture Notes in Computer Science, vol 11690

Experiential AI

Drew Hemment, Ruth Aylett, Vaishak Belle, Dave Murray-Rust, Ewa Luger, Jane Hillston, Michael Rovatsos, Frank Broz

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI、cs.LG

Comments To appear in AI Matters 5(1): 25-31 (2019)

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Chandra Khatri, Behnam Hedayatnia, Anu Venkatesh, Jeff Nunn, Yi Pan, Qing Liu, Han Song, Anna Gottardi, Sanjeev Kwatra, Sanju Pancholi, Ming Cheng, Qinglang Chen, Lauren Stubel, Karthik Gopalakrishnan, Kate Bland, Raefer Gabriel, Arindam Mandal, Dilek Hakkani-Tur, Gene Hwang, Nate Michel, Eric King, Rohit Prasad

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Comments 2018 Alexa Prize Proceedings

Relational Neural Expectation Maximization: Unsupervised Discovery of Objects and their Interactions

Sjoerd van Steenkiste, Michael Chang, Klaus Greff, Jürgen Schmidhuber

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI、cs.LG

Comments Accepted to ICLR 2018

Cognitive Database: A Step towards Endowing Relational Databases with Artificial Intelligence Capabilities

Rajesh Bordawekar, Bortik Bandyopadhyay, Oded Shmueli

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Using NLU in Context for Question Answering: Improving on Facebook's bAbI Tasks

John S. Ball

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Comments 38 Pages, 10 Tables

Deal or No Deal? End-to-End Learning for Negotiation Dialogues

Mike Lewis, Denis Yarats, Yann N. Dauphin, Devi Parikh, Dhruv Batra

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Universal Memory Architectures for Autonomous Machines

Dan P. Guralnik, Daniel E. Koditschek

专题命中 复杂问题求解 ：planning(abstract)；分类 cs.AI、cs.LG

Comments Technical report, 31 pages, 1 table, 14 figures, 2 appendices

Mixing Metaphors

Mark Lee, John Barnden

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL、cs.AI

Journal ref Proceedings of the AISB'99 Symposium on Metaphor, Artificial Intelligence, and Cognition, pages 11-16, Edinburgh

A Cortically Inspired Architecture for Modular Perceptual AI

一种模块化感知人工智能的皮层启发架构

Prerna Luthra

机构 * Independent Researcher（独立研究者） 纠错

专题命中 复杂问题求解 ：reasoning(abstract,comments)；分类 cs.AI

AI总结 本文提出了一种受皮层启发的模块化感知人工智能架构，通过分解感知为专门模块，提升可解释性和推理透明度。

Comments Accepted to the ICLR 2026 Workshop on "From Human Cognition to AI Reasoning: Models, Methods, and Applications (HCAIR)"

Tree-of-Code: A Tree-Structured Exploring Framework for End-to-End Code Generation and Execution in Complex Task Handling

Ziyi Ni, Yifan Li, Ning Yang, Dou Shen, Pin Lv, Daxiang Dong

机构 * The Key Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences（认知与决策智能复杂系统重点实验室，自动化研究所，中国科学院） ； School of Artificial Intelligence, University of Chinese Academy of Science（人工智能学院，中国科学院大学） ； Baidu, Inc.（百度公司） ； Global Innovation Exchange Institution, Tsinghua University（全球创新交换机构，清华大学） 纠错

专题命中 复杂问题求解 ：reasoning(abstract,comments)；分类 cs.AI

Comments This idea was first submitted to the NeuralPS Workshop "System 2 Reasoning At Scale" in September 2024. Its OpenReview: https://openreview.net/forum?id=8NKAL8Ngxk&noteId=8NKAL8Ngxk. It was then submitted to the NAACL 2025 in October 2024, which is recorded in: https://openreview.net/forum?id=S0ZUWD3Vy5&noteId=S0ZUWD3Vy5. Now this paper has been accepted for publication in ACL 2025 Findings

Compositional Chain-of-Relations for Faithful Knowledge Graph Question Answering with Large Language Models

面向大型语言模型的忠实知识图谱问答的组合式关系链

Chenhui Liu, Jianpeng Zhou, Jiahai Wang

机构 * Sun Yat-sen University（中山大学） 纠错

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

AI总结 本文针对现有基于LLM的KGQA方法存在的实体剪枝不可靠、约束处理无基础的局限，提出组合式关系链（CCoR）框架，以关系为中心探索实现更忠实的多跳KGQA，在多个基准上性能优于基线。

Think with Structured Grounding: Perceptual Reinforcement Learning for Chart and Visual-Tabular Understanding

基于结构化接地的思考：面向图表与视觉表格理解的感知强化学习

Changjiang Jiang, Qiannian Zhao, Lei Xin, Jinxiang Xie, Preslav Nakov, Zhuohan Xie

机构 * Mohamed bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学） ； Nanjing University（南京大学） 纠错

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

AI总结 针对MLLMs依赖外部工具感知图表等视觉内容的问题，提出TwSG框架，通过两阶段训练实现细粒度视觉推理，降低延迟并提升准确率与鲁棒性。

Comments Manuscript

PropUQ-MAS: Propagation-Aware Uncertainty Quantification for LLM Multi-Agent Systems

PropUQ-MAS：面向大语言模型多智能体系统的传播感知不确定性量化

Yaokun Liu, Yifan Liu, Daniel Yue Zhang, Ruichen Yao, Zelin Li, Dong Wang

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL

AI总结 针对现有不确定性量化方法无法捕捉大语言模型多智能体系统中不确定性传播的问题，提出PropUQ-MAS框架，实验显示其可显著提升该系统的不确定性量化性能。

Comments Accepted to EMNLP 2026 (Main Conference)

GenCoord: Skill-Path Commitments under Private Information

Peng He, Junning Zhu, Haohan Yuan, Jianpeng Liang

机构 * Tsinghua University（清华大学） ； Beijing Normal-Hong Kong Baptist University（北京师范大学-香港浸会大学联合国际学院） ； University of North Carolina at Charlotte（北卡罗来纳大学夏洛特分校） ； University of California San Diego（加利福尼亚大学圣迭戈分校） 纠错

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

Comments 25 pages, 10 figures, and 23 tables, including supplementary material. Peng He and Junning Zhu contributed equally. Paper source and compact evidence: this https URL (https://github.com/JulianZJN/GenCoord)

Hierarchical Compositionality for An Assistive AI Agent

面向辅助AI智能体的分层组合性

Tianyi Fu, Mohan Sridharan

机构 * University of Edinburgh（爱丁堡大学） 纠错

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

AI总结 本文针对辅助AI智能体的歧义问题，提出嵌入分层组合性原则的架构，结合语义兼容性等模型推理实现歧义消除，实验表明其性能优于当前最优数据驱动基线，可适配特定用户画像。

Comments 29 pages, 9 figures, 4 tables. Project page: this https URL (https://tianyi-fu.github.io/HCAA)

SDAD: Spec-Driven Agentic Development for the AI-Native SDLC

SDAD：面向AI原生软件开发生命周期的规范驱动智能体开发

Vu Hung Nguyen, Thanh Nguyen

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

AI总结 该研究提出规范驱动智能体开发（SDAD）模型，对比传统敏捷开发，扩展团队角色、量化治理等，论证智能体开发需将工程规范转移至上游规范环节。

Tree-of-Experience: Hierarchical Experience Management for Self-Evolving Agents

经验树：面向自我进化智能体的分层经验管理

Zihao Deng, Yining Zhu, Leiming Wang, Junbo Wang, Jingfei Lu

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.CL

AI总结 该研究提出经验树（ToE）框架，将经验组织与LLM智能体分层推理对齐，在“24点游戏”和“金融进化基准”上大幅提升了解题性能与效率，解决了现有经验表示与推理过程脱节的问题。

Weak Critics Make Strong Learners: On-Policy Critique Distillation for Scalable Oversight

弱批评者造就强学习者：用于可扩展监督的在策略批评蒸馏

Can Jin, Jiakang Li, Rui Wu, Eddy Z. Zhang, Dimitris N. Metaxas

机构 * University of Cambridge（剑桥大学） ； University of California, Berkeley（加州大学伯克利分校） ； UC Berkeley AI Lab（加州大学伯克利分校人工智能实验室） 纠错

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

AI总结 提出在策略批评蒸馏（OPCD）方法，利用弱模型作为批评者提供修订方向，通过自适应自教师信号蒸馏批评引导的行为，提升强模型在推理和对齐基准上的表现。

FlowEvo: Self-Evolving Agents through the Co-Evolution of Workflows and Executable Skills

FlowEvo：通过工作流和可执行技能的协同进化实现自我进化的智能体

Zeyu Ren, Ling Yue, Ran Li, Yishu Wang, Shengxiang Xu, Hanmo Liu, Shaowu Pan, Shimin Di

机构 * Southeast University（东南大学） ； Rensselaer Polytechnic Institute（伦斯勒理工学院） ； The Hong Kong University of Science and Technology（香港科技大学） 纠错

专题命中 复杂问题求解 ：reasoning(abstract)；分类 cs.AI

AI总结 研究旨在让大型语言模型智能体通过工作流解决复杂任务。提出FlowEvo框架，通过工作流到技能编译、技能到工作流反馈、技能管理三个机制，使智能体无需更新模型参数积累完善能力，实验显示其在基准测试中精度-成本权衡优，各机制有贡献。

Comments Published as a conference paper at the Conference on Language Modeling (COLM) 2026. 25 pages, 3 figures, 16 tables. Code: https://github.com/DEFENSE-SEU/FlowEvo