Reasoning over natural language is a long-standing goal for the research community. However, studies have shown that existing language models are inadequate in reasoning. To address the issue, we present POET, a novel reasoning pre-training paradigm. Through pre-training language models with programs and their execution results, POET empowers language models to harvest the reasoning knowledge possessed by program executors via a data-driven approach. POET is conceptually simple and can be instantiated by different kinds of program executors. In this paper, we showcase two simple instances POET-Math and POET-Logic, in addition to a complex instance, POET-SQL. Experimental results on six benchmarks demonstrate that POET can significantly boost model performance in natural language reasoning, such as numerical reasoning, logical reasoning, and multi-hop reasoning. POET opens a new gate on reasoning-enhancement pre-training, and we hope our analysis would shed light on the future research of reasoning like program executors.
StrategyLLM: Large Language Models as Strategy Generators, Executors, Optimizers, and Evaluators for Problem Solving
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University(上海交通大学电子信息与电气工程学院);The Chinese University of Hong Kong(香港中文大学);Tencent AI Lab(腾讯AI实验室)
SURGE: On the Potential of Large Language Models as General-Purpose Surrogate Code Executors
Department of Computer Science and Technology, Tsinghua(清华大学计算机科学与技术系);Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua(清华大学交叉信息研究院)
Large Language Models as Code Executors: An Exploratory Study
Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Xinjiang University(新疆大学);University of Melbourne(墨尔本大学);Tsinghua University(清华大学);Shanghai Jiao Tong University(上海交通大学);IBM Research(IBM研究院);Nantong University(南通大学);Alibaba(阿里巴巴)
TITAN: Graph-Executable Reasoning for Cyber Threat Intelligence
Sapienza Università di Roma(萨皮恩扎罗马大学);Institute of Informatics and Telematics (IIT)(信息与电信学研究院);National Research Council of Italy (CNR)(意大利国家研究委员会);Department of Excellence in Robotics and AI(机器人与人工智能卓越部门);TeCIP Institute(TeCIP研究所);Scuola Superiore Sant'Anna(圣安娜高级学校)
Fact-Checking Complex Claims with Program-Guided Reasoning
University of California, Santa Barbara(加州大学圣巴巴拉分校);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Nanyang Technological University(南洋理工大学);National University of Singapore(新加坡国立大学)
Grounding Physical Concepts of Objects and Events Through Dynamic Visual Reasoning
The University of Hong Kong(香港大学);MIT CSAIL(麻省理工学院计算机科学与人工智能实验室);Stanford University(斯坦福大学);MIT BCS, CBMM, CSAIL(麻省理工学院脑与认知科学系、大脑、心智与机器中心、计算机科学与人工智能实验室);MIT-IBM Watson AI Lab(MIT-IBM沃森人工智能实验室)
KG-Agent: An Efficient Autonomous Agent Framework for Complex Reasoning over Knowledge Graph
Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学高瓴人工智能学院);School of Information, Renmin University of China(中国人民大学信息学院);NLP Center, BOSS Zhipin(BOSS直聘NLP中心);Career Science Lab, BOSS Zhipin(BOSS直聘职业科学实验室)
Improved Logical Reasoning of Language Models via Differentiable Symbolic Programming
Carnegie Mellon University(卡内基梅隆大学);University of Pennsylvania(宾夕法尼亚大学);Mohamed Bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Petuum Inc.(Petuum公司)
PRIME: Policy-Reinforced Iterative Multi-agent Execution for Algorithmic Reasoning in Large Language Models
Purdue University(普渡大学);University of Malaya(马来亚大学);Faculty of Information Technology, Beijing University of Technology(北京理工大学信息学院);School of Computer Science, Georgia Institute of Technology(佐治亚理工学院计算机科学学院);Department of Industrial and Systems Engineering, Lehigh University(莱斯大学工业与系统工程系)
Towards Better Understanding of Program-of-Thought Reasoning in Cross-Lingual and Multilingual Environments
School of Information Science and Technology, VISTEC(信息科学与技术学院,VISTEC);KAIST(韩国科学技术院);Cohere;SCB 10X;AI Singapore(AI新加坡);Department of Computer Engineering, Chulalongkorn University(朱拉隆梭大学计算机工程系)
MultiLingPoT: Enhancing Mathematical Reasoning with Multilingual Program Fine-tuning
Fudan University(复旦大学);MYbank, Ant Group(蚂蚁集团MYbank);School of Computer Science, Fudan University(复旦大学计算机学院);School of Data Science, Fudan University(复旦大学数据科学学院)