AI-Enhanced Sustainable Energy Management and Policy Recommendations for the U.S. Power Sector
Bibliographic record
Abstract
The transition to a sustainable U.S. power sector is faced with the challenge of balancing reliability, decarbonization, and affordability in more variable demand conditions. There is tremendous potential for artificial intelligence (AI) to make huge contributions to forecasting accuracy, optimizing resource allocation, and in developing policy and evidence-based policy. This paper presents a new model of hybrid artificial intelligence for sustainable energy management applications that consists of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU) and Transformer networks combined with CatBoost meta-learner and then optimized with residual BiGRU-attention and α-blender. The proposed model was tested using the public dataset of hourly electricity generation and demand data from the United States of America with temporal and operational characteristics significant to grid planning. Experimental results indicate a very high performance of the hybrid model compared with baseline methods for all performance measures. Specifically, it had a minimum MAE (0.007503), minimum MSE (0.000102) and RMSE (0.010082) and highest R2 value (0.995954) than single Lstm, GRU, Catboost and Xgboost. Aside from technical performance, the study accomplishes three contributions: efficiency-enhancing energy management using AI-driven energy forecasting and optimization, American context - use of demand forecasting for renewable integration and grid balancing, and extending technical understanding to the policy advice for future adaptive regulations. By integrating AI analytics, policy design, and sustainability policy, this research makes a methodological contribution as well as a governance-centric framework with a focus on setting up AI as an accelerator for the sustainable transformation of America's power grid.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 source (direct Gemma or distilled Codex), 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".