EF-LLM: Energy Forecasting LLM with AI-assisted Automation, Enhanced Sparse Prediction, Hallucination Detection

TL;DR

EF-LLM tackles the need for end-to-end energy forecasting with AI-assisted automation, robust performance under data sparsity, and transparent hallucination detection. It integrates a multi-channel architecture that fuses time-series and textual knowledge, a loss framework L = L_orig + lambda * ||DeltaW||^2 with L_orig = varpi L1 + (1 - varpi) L2, and a Fusion Parameter-Efficient Fine-Tuning (F-PEFT) that combines Prefix Tuning and LoRA, along with function calling. Continual learning is enabled via updatable LoRA injections, and heterogeneous multimodal data alignment ensures coherent knowledge transfer across tasks like load, PV, and wind forecasting. Hallucination analysis combines semantic similarity measures with ANOVA to quantify reliability, demonstrating strong performance in sparse-data regimes and reliable outputs essential for energy decision-making. Collectively, EF-LLM advances toward fully automated, end-to-end energy forecasting with interpretable reliability, making it practically impactful for grid operations and planning.

Abstract

Accurate prediction helps to achieve supply-demand balance in energy systems, supporting decision-making and scheduling. Traditional models, lacking AI-assisted automation, rely on experts, incur high costs, and struggle with sparse data prediction. To address these challenges, we propose the Energy Forecasting Large Language Model (EF-LLM), which integrates domain knowledge and temporal data for time-series forecasting, supporting both pre-forecast operations and post-forecast decision-support. EF-LLM's human-AI interaction capabilities lower the entry barrier in forecasting tasks, reducing the need for extra expert involvement. To achieve this, we propose a continual learning approach with updatable LoRA and a multi-channel architecture for aligning heterogeneous multimodal data, enabling EF-LLM to continually learn heterogeneous multimodal knowledge. In addition, EF-LLM enables accurate predictions under sparse data conditions through its ability to process multimodal data. We propose Fusion Parameter-Efficient Fine-Tuning (F-PEFT) method to effectively leverage both time-series data and text for this purpose. EF-LLM is also the first energy-specific LLM to detect hallucinations and quantify their occurrence rate, achieved via multi-task learning, semantic similarity analysis, and ANOVA. We have achieved success in energy prediction scenarios for load, photovoltaic, and wind power forecast.

Figures (18)

• Figure 1: EF-LLM Prefix Tuning Block.
• Figure 2: EF-LLM text preprocessing layer.
• Figure 3: EF-LLM Deep Network Architecture with LoRA.
• Figure 4: Heterogeneous Data Alignment: Multi-Channel Architecture.
• Figure 5: Heterogenous Multimodal Data Alignment.