A Deep Reinforcement Learning Based Dynamic False Data Injection Detection in Load Prediction
Bibliographic record
Abstract
The integration of smart meters into modern energy distribution systems has improved grid efficiency and consumer energy management. However, it also introduces new cybersecurity vulnerabilities, particularly False Data Injection (FDI) attacks that can manipulate energy consumption data. Traditional anomaly detection approaches often rely on supervised learning and require extensive labeled attack data, which may not generalize well to evolving cyber-threats. To address these limitations, this paper formulates the FDI detection problem as a partially observable Markov decision process (POMDP) and leverages the proximal policy optimization (PPO) reinforcement learning (RL) algorithm to solve the problem. A belief estimation model is developed using a multi-head self-attention architecture to represent the agent's uncertainty about whether the data point is attacked. The belief model, as a supervised learning model, is also used for comparing the results which proves the superior performance of the PPO-RL agent in detecting sequential and random scaling attacks at varying intensities. The PPO-RL agent also gains a higher reward than an RL agent trained based on the Advantage Actor-Critic (A2C) algorithm.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".