Apr 22
From Natural Language to Executable Narsese: A Neuro-Symbolic Benchmark and Pipeline for Reasoning with NARS
★★★★★
significance 3/5
This paper introduces a neuro-symbolic framework that translates natural language reasoning into executable Narsese code. It also presents the NARS-Reasoning-v0.1 benchmark, which includes a compilation pipeline and a Phi-2 LoRA adapter for improved symbolic reasoning.
Why it matters
Bridging the gap between linguistic fluidity and formal logic remains critical for building reliable, verifiable reasoning architectures in autonomous agents.
Tags
#neuro-symbolic #llm reasoning #nars #benchmark #formal logicRelated coverage
- Global South OpportunitiesPivotal Research Fellowship 2026 (Q3): AI Safety Research Opportunity - Global South Opportunities
- arXiv cs.AIAn Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on Multi-Fidelity Digital Twin and FMEA Knowledge Enhancement
- arXiv cs.AIPExA: Parallel Exploration Agent for Complex Text-to-SQL
- arXiv cs.AIThe Power of Power Law: Asymmetry Enables Compositional Reasoning
- arXiv cs.AIOn the Existence of an Inverse Solution for Preference-Based Reductions in Argumentation