MinaGabriel/NARS-Reasoning-v0.1

NARS-Reasoning-v0.1是一个结合了第一顺序逻辑(FOL)和非公理化推理系统(NARS)的正式表示语言Narsese的符号推理数据集。它由ProverQA数据集衍生而来，包含了经过符号证明者验证的高质量FOL推理数据，并重新编码为Narsese以确保与OpenNARS和其他认知推理引擎的兼容性。数据集可用于自然语言到Narsese翻译、混合神经符号推理、逻辑准确性和不确定性的评估以及推理系统的基准测试。NARS-Reasoning-v0.1 is a symbolic reasoning dataset that combines First-Order Logic (FOL) with Narsese, the formal representation language of the Non-Axiomatic Reasoning System (NARS). It is derived from the ProverQA dataset and contains high-quality FOL reasoning data validated by symbolic provers, re-encoded into Narsese for compatibility with OpenNARS and other cognitive reasoning engines. The dataset supports natural-language to Narsese translation, hybrid neuro-symbolic reasoning, evaluation of logical accuracy and uncertainty, and benchmarking reasoning systems.