Research map / Time series / Foundation models for time series
Satellite-Text-Prompted Large Language Model for Photovoltaic Power Forecasting
Pengfei Jia, Jianghong Ma, Baoquan Zhang, Kenghong Lin, Xinyu Zhang, Chuyao Luo, Xutao Li, Yunming Ye
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Semantic Adapter (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04750 -> 0.03962 | Improved |
| add pretrained LLM backbone (ablation) | PV-LLM · MMSP(S) All | MAE: 0.05500 -> 0.03962 | Improved |
| use pretrained weights instead of training the LLM from scratch (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04610 -> 0.03962 | Improved |
| freeze the LLM (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04209 -> 0.03962 | Improved |
| zero-initialize adaptive gate (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04202 -> 0.03962 | Improved |
| add Text Prompt Construction (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04255 -> 0.03962 | Improved |
| add Adaptive Prompt Tuning layers (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04405 -> 0.03962 | Improved |
| add Context Compressor (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04376 -> 0.03962 | Improved |
| add forecast start time to text prompt (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04314 -> 0.03962 | Improved |
| add plant coordinates to text prompt (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04103 -> 0.03962 | Improved |
| add dataset/task descriptions and input statistics to prompt (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04027 -> 0.03962 | Improved |
| use RoPE instead of sinusoidal/learned positional encoding (ablation) | PV-LLM · MMSP(S) All | MAE: 0.04070 (sinusoidal), 0.04152 (learned) -> 0.03962 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Time-LLM — Time-LLM: Time Series Forecasting by Reprogramming Large Language Models (ICLR 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Solar Power Generation Forecasting via Multimodal Feature Fusion (Student Abstract) (AAAI 2024 · similarity 0.87)
- $S^2$IP-LLM: Semantic Space Informed Prompt Learning with LLM for Time Series Forecasting (ICML 2024 · similarity 0.86)
- Time-LLM: Time Series Forecasting by Reprogramming Large Language Models (ICLR 2024 · similarity 0.85)
- GPT4MTS: Prompt-based Large Language Model for Multimodal Time-series Forecasting (AAAI 2024 · similarity 0.85)
- CALF: Aligning LLMs for Time Series Forecasting via Cross-modal Fine-Tuning (AAAI 2025 · similarity 0.85)
- TimeCMA: Towards LLM-Empowered Multivariate Time Series Forecasting via Cross-Modality Alignment (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
