Hybrid NARX-LLM for Greenland Iceberg Discharge: Prompt-Driven Residual Correction
Researchers have developed a novel Hybrid NARX-LLM framework to improve predictions of Greenland iceberg discharge. This approach combines a nonlinear autoregressive model (NARX) with a large language model (LLM) for residual correction, enhanced by a Physics-Informed Prompt (PIP) method. The PIP transforms physical knowledge into structured prompts, enabling the LLM to reason about unmodeled factors and correct systematic prediction errors, particularly for extreme events. AI