Addressing Unintended Bias in Toxicity Detection: An LSTM and Attention-Based Approach
Bibliographic record
Abstract
In the digital era, online platforms serve as crucial hubs for social interactions and idea exchange. However, these platforms are continually shadowed by toxic comments that undermine genuine discourse and have the potential to harm participants. While machine learning provides an avenue for detecting such toxic content, a significant challenge arises when these models, influenced by biased training datasets, inadvertently propagate or amplify inherent biases. Such unintentional biases are especially disconcerting when they disadvantage or misrepresent identities already vulnerable in online spaces. Addressing this complex landscape, our research presents a model meticulously designed to detect toxic comments, aiming to achieve a higher degree of accuracy while striving to minimize such unintended biases. Our approach is underpinned by a combination of a tailored data preprocessing technique and the integration of Long Short-Term Memory networks (LSTM) with Attention mechanisms. Preliminary evaluations reveal our model's AVC score to be 0.93524, indicating its efficacy in toxicity detection. While there's always room for improvement, the design and results of our model emphasize the importance and feasibility of developing more nuanced and unbiased machine learning solutions for the challenges posed in the digital domain.
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 | 0.002 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".