Research on HSE Risk Assessment Method for Multi-source Heterogeneous Data Driven by Transformer-FL Framework
Bibliographic record
Abstract
With the widespread application of multi-source heterogeneous data in HSE risk assessment, how to efficiently integrate different data sources and improve assessment accuracy has become an urgent problem to be solved. This paper proposes an HSE risk assessment method based on Transformer and federated learning, aiming to improve the accuracy of risk prediction through the effective integration of multi-source data. In this method, the Transformer model captures deep correlations in multi-source data through self-attention mechanism, and federated learning ensures cross-device collaborative training while protecting data privacy. Experimental results using a multi-source heterogeneous dataset from the chemical, manufacturing, and energy industries show that the Transformer-FL framework significantly improves risk assessment accuracy. The dataset includes real-time environmental data, accident records, and operation logs. Experiments on high-performance workstations with Nvidia RTX A6000 GPUs and Intel Xeon processors reported accuracy improvements: chemical industry (58.9% to 41.2%), manufacturing (35.6% to 23.4%), and energy industry (50.1% to 36.8%). The Transformer-FL framework has reduced the HSE risk value of traditional methods from 58.9% to 41.2%, indicating a lower risk, while the accuracy of risk assessment has improved by 17.7%. It is important to note that the percentages in this context refer to the risk value, where a lower value signifies reduced risk, and the accuracy improvement refers to the increase in correct predictions. It is important to note that the percentages in this context refer to the risk level; a lower percentage represents a lower residual risk, which indicates improvement. In contrast, accuracy improvements are calculated as the percentage increase in correct predictions. In the manufacturing industry, despite strong data homogeneity, the accuracy rate has increased from 35.6% to 23.4%, demonstrating the advantages of this framework in heterogeneous data environments. The experimental results show that the Transformer-FL framework has significant advantages in different HSE scenarios, especially when the amount of data is large, and the fusion effect far exceeds that of traditional risk assessment methods. Overall, this framework provides an intelligent, efficient and privacy-protected solution for HSE risk assessment, which can meet the dual needs of multi-source heterogeneous data processing and security in the industrial field.
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 | 0.004 | 0.008 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.004 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| 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".