IoTDL<sup>2</sup>AIDS: Toward IoT-Based System Architecture Supporting Distributed LSTM Learning for Adaptive IDS on UAS
Bibliographic record
Abstract
The rapid proliferation of Unmanned Aircraft Systems (UAS) introduces new threats to national security. UAS technologies have dramatically revolutionized legitimate business operations while providing powerful weaponizing systems to malicious actors and criminals. Due to their inherited wireless capabilities, they are an easy target for cyber threats. In response to this challenge, the implementation of many Intrusion Detection Systems (IDS), which support anomaly detection on UAS, have been proposed in the past. However, such systems often require offline training with heavy processing, making them unsuitable for UAS deployment. This is pertinent for drone systems that support dynamic changes in mission operational tasks. This paper presents a novel system architecture that utilizes sensing systems capabilities available on existing IoT infrastructure for supporting rapid infield adaptive models’ training and parameters estimation services for UAS. We have devised a cluster-oriented distributed training algorithm based on LSTM with mini-batch gradient descent, with hundreds of IoT platforms per cluster collaboratively performing model parameters estimation tasks. The proposed architecture is based on deploying a multilayer system that facilitates secure dissemination of power consumption behavioral patterns for the flight sensing system between the UAS layer and the IoT layer. The model was implemented and deployed on a real IoT-enabled platform based on NXP-Kinetis K64–120 MHz. Furthermore, model training and validation were performed by applying various datasets contaminated with different percentages of malicious data. Our anomaly detection model achieved high prediction accuracy with an ROC-AUC score of 0.9332. The model maintains minimal power consumption overheads and low training time during the processing of a data batch.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.002 |
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".