On the Benefits of Transfer Learning and Reinforcement Learning for Electric Short-term Load Forecasting
Bibliographic record
Abstract
Accurate short-term load forecasting plays an essential role in effective modern power system operations. Recently, various deep learning based time-series forecasting algorithms have shown superior performances. Usually, the existing time-series forecasting algorithms require an adequate amount of training samples to learn a reliable prediction model. However, in some real-world scenarios, we might only have limited training samples, i.e., learning to predict electric load in a newly built neighborhood. Under such strict constraints, both classical and deep learning based time-series forecasting algorithms suffer from high prediction errors and over-fitting problems due to the limited training data. On the other hand, in the real world, we may have a large amount of historical data collected from other buildings which could be helpful to learn the forecasting model. Therefore, in this work, we propose to tackle the short-term residential electric load forecasting problem from a transfer learning perspective. The goal is to use the large amount of historical data from other source buildings to learn reliable forecasting models for the target building which only has limited training data. In particular, we first use the Autoformer, a state-of-the-art (SOTA) transformer-based time-series forecasting algorithm, to learn a forecasting model from each source building, respectively. Then, we leverage the benefit of the reinforcement learning algorithm to select the learned forecasting models to make prediction on the target building, named Time-Series Double DQN (TS-DDQN). To validate the efficacy of the proposed method, we conduct extensive experiments on different real-world datasets. Experimental results show that TS-DDQN can consistently outperform baseline algorithms by a large margin.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".