An Integration of Fine-Tuned RoBERTa, Support Vector Machine, and Differential Evolution in Enhancing Multi-Label Text Classification for Academic Publications
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Two important challenges in multi-label classification are dataset imbalance and overlapping features shared among two or more class labels. This study addresses both challenges as a problem of classifying multi-labels in academic publications by proposing an integration of three main techniques: Robustly Optimized BERT Approach (RoBERTa), Support Vector Machine (SVM), and Differential Evolution (DE) algorithm. In a nutshell, a fine-tuned RoBERTa captures domain-specific semantic features of academic papers; SVM performs initial multi-label classification; and DE optimizes the label-specific threshold combination for each label improving the accuracy of classification decisions. This integration jointly addresses dataset imbalance, overlapping features, and threshold selection issues, which are rarely considered together in prior studies. The experiments were performed on academic papers, collected from arXiv repository with five core subfields of Computer Science. Experimental results showed that the Macro-F1 score of the proposed approach that integrates the fine-tuned RoBERTa model, SVM classifier and DE achieved 0.8018 while the best performance of benchmark methods achieved 0.7218. The significant performance improvement highlights the effectiveness of integrating semantic feature extraction, robust classification, and adaptive threshold optimization to overcome the challenges of dataset imbalance and feature overlap in multi-label classification of academic publications.
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.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.002 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.000 |
| 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 it