Reasoning System
Reasoning systems aim to build artificial intelligence capable of performing complex logical deductions and decision-making, mirroring human cognitive abilities. Current research emphasizes improving the reliability and efficiency of these systems, focusing on hybrid architectures that combine neural networks with symbolic reasoning methods, such as those employing chain-of-thought prompting, knowledge graphs, and various forms of logic-based reasoning. These advancements are crucial for enhancing the trustworthiness and explainability of AI systems across diverse applications, from medical diagnosis and legal reasoning to scientific discovery and robotics.
Papers
Aristotle: Mastering Logical Reasoning with A Logic-Complete Decompose-Search-Resolve Framework
Jundong Xu, Hao Fei, Meng Luo, Qian Liu, Liangming Pan, William Yang Wang, Preslav Nakov, Mong-Li Lee, Wynne Hsu
OpenRFT: Adapting Reasoning Foundation Model for Domain-specific Tasks with Reinforcement Fine-Tuning
Yuxiang Zhang, Yuqi Yang, Jiangming Shu, Yuhang Wang, Jinlin Xiao, Jitao Sang
Imitate, Explore, and Self-Improve: A Reproduction Report on Slow-thinking Reasoning Systems
Yingqian Min, Zhipeng Chen, Jinhao Jiang, Jie Chen, Jia Deng, Yiwen Hu, Yiru Tang, Jiapeng Wang, Xiaoxue Cheng, Huatong Song, Wayne Xin Zhao, Zheng Liu, Zhongyuan Wang, Ji-Rong Wen
A NotSo Simple Way to Beat Simple Bench
Soham Sane, Angus McLean