Skip to content

📡 Signal & Communications

💬 ACL2025 · 1 paper notes

📌 Same area in other venues: 📷 CVPR2026 (2) · 🔬 ICLR2026 (8) · 🧪 ICML2026 (2) · 🤖 AAAI2026 (3) · 🧠 NeurIPS2025 (5) · 📹 ICCV2025 (3)

WirelessMathBench: A Mathematical Modeling Benchmark for LLMs in Wireless Communications

This paper presents WirelessMathBench, a mathematical modeling benchmark for wireless communications featuring 587 problems extracted from 40 cutting-edge papers. It systematically evaluates the capabilities of LLMs in domain-specific mathematical derivations, revealing that even the strongest model, DeepSeek-R1, achieves an average accuracy of only 38.05%, and a mere 7.83% in full formula derivation.