Recall-First Moderation via Distribution-Preserving Augmentation and Committee-Diverse Retrieval

Alejandro Salinas-Medina, Xue LiuOpenReview 2026 · Submission to ICLR 2026 · Rejected · OpenReview

Abstract

False negatives—missed unsafe content—remain the dominant risk in safety-critical moderation. We present a novel recall-first moderation framework that integrates two complementary innovations: (i) distribution-preserving contrastiveaugmentation, which generates boundary-focused hard positives and negatives while statistically preserving corpus structure, and (ii) committee-diverse re-trieval, which combines dense, MMR, and graph-based selectors to construct label-informative, non-redundant neighborhoods at inference. Augmented corpora are validated with KL/JS divergence thresholds (≤ 0.05 globally), confirming indistinguishability from the source distribution. On a large held-out test se tof multidomain unbalanced text, vanilla retrieval-augmented pipelines expose the persistent failure mode of under-detecting FLAGGED content (recall ≈ 0.44), but also reveal a strong baseline gap: an open-source stack (FAISS + local LLaMA-3) achieves significantly higher accuracy and macro-F1 than a commercial counter-part (API embeddings + hosted LLM). Adding augmentation and committee retrieval improves sensitive-class recall by ∼ 10 points (to ≈ 0.56) while maintaining global performance, with graph-aware retrieval pushing open-source accuracy to 0.8510 and Macro-F1 to 0.7635. Ensemble experiments with DistilRoBERTa further raise recall to 0.5781 without loss of utility.
BibTeX
@misc{salinasmedina2026recall,
  author = {Alejandro Salinas-Medina and Xue Liu},
  title = {{Recall-First Moderation via Distribution-Preserving Augmentation and Committee-Diverse Retrieval}},
  year = {2026},
  url = {https://openreview.net/forum?id=qfIMtZQspu},
  howpublished = {OpenReview},
  note = {Submitted to ICLR 2026},
}

Loading related content...

Figure