Short Text Classification Based on Hybrid Semantic Expansion and Bidirectional GRU (BiGRU) based Method to Improve Hate Speech Detection
Bibliographic record
Abstract
The persistent prevalence of hate speech on contemporary social media platforms demands advanced detection methods to address specific categories and levels of offenses.This research focuses on enhancing hate speech detection by refining text representation through a semantic expansion approach, surpassing the limitations of conventional methods.The back-translation technique is employed to enhance sentence structure.Initially, the Lesk Algorithm is utilized for word disambiguation in the semantic expansion process, identifying word meanings within relevant contexts.Subsequently, knowledge bases from WordNet and Kateglo are leveraged to enrich contextual information.The final step involves using Cosine Similarity to select the most appropriate words based on the highest scores.The combined semantic expansion technique significantly improves classification performance compared to conventional methods.Data, with and without semantic expansion, is vectorized into the BERT embedding space and classified using deep learning models such as CNN, BiGRU, and BiLSTM.The proposed approach consistently demonstrates high accuracy across all model types: CNN (88%), BiGRU (88.3%), and BiLSTM (87.3%).In contrast, models without semantic expansion yield relatively lower results-CNN (83.6%),BiGRU (83.3%), and BiLSTM (83.1%).This underscores the substantial breakthrough of the semantic expansion approach in overcoming challenges related to data distribution and semantic feature scarcity, ultimately resulting in improved classification performance.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.003 |
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".