🚗 Autonomous Driving

💬 ACL2025 · 1 paper notes

📌 Same area in other venues: 📷 CVPR2026 (157) · 🔬 ICLR2026 (50) · 🧪 ICML2026 (8) · 🤖 AAAI2026 (56) · 🧠 NeurIPS2025 (47) · 📹 ICCV2025 (91)

Embracing Large Language Models in Traffic Flow Forecasting

The LEAF framework is proposed, which utilizes a dual-branch predictor comprising a graph branch (for pair-wise relations) and a hypergraph branch (for non-pair-wise relations) to generate candidate forecasts. A frozen LLM is then employed as a selector (interpreting discriminatively rather than generatively) to choose the optimal forecast, optimizing the predictor through feedback via ranking loss, achieving SOTA on PEMS datasets.