Multi-Modal Personalization for Toxicity Detection via Reinforcement-Learned Spiking Neural Fusion
Bibliographic record
Abstract
Toxicity detection in online interactions is a critical challenge for fostering safe and inclusive digital environments. Traditional approaches often rely on single-modality data or static models, limiting their ability to capture nuanced and context-dependent toxic behaviors. This paper presents a novel multi-modal framework for personalized toxicity detection, leveraging Convolutional Spiking Neural Networks (CSNNs) and reinforcement learning (RL). The system integrates text, audio, and visual modalities through a late fusion strategy, enabling the detection of complex toxic cues spanning verbal, paralinguistic, and non-verbal domains. Reinforcement learning dynamically adapts the model to user-specific preferences, addressing cultural and contextual variations in interpreting toxicity. Experimental results demonstrate the framework's effectiveness, achieving superior performance and significantly outperforming state-of-the-art models. The integration of attention mechanisms and spiking neurons enhances interpretability and temporal modeling, while RL-driven fine-tuning reduces false positives and negatives, improving user satisfaction. This work underscores the potential of combining multimodal learning and reinforcement learning for robust and adaptable toxicity detection. Future directions include expanding datasets to capture diverse cultural contexts, optimizing computational efficiency for real-time applications, and exploring additional modalities to further enhance performance. The proposed framework represents a significant step toward intelligent, user-centric solutions for toxicity detection across diverse digital platforms.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| 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) | 0.001 | 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".