Dynamic Content Moderation in Livestreams: Combining Supervised Classification with MLLM-Boosted Similarity Matching

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yew, Wei Chee, Xu, Hailun, Saha, Sanjay, Fan, Xiaotian, Ong, Hiok Hian, Wang, David Yuchen, Sarkar, Kanchan, Yang, Zhenheng, Guan, Danhui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908792944328704
author Yew, Wei Chee
Xu, Hailun
Saha, Sanjay
Fan, Xiaotian
Ong, Hiok Hian
Wang, David Yuchen
Sarkar, Kanchan
Yang, Zhenheng
Guan, Danhui
author_facet Yew, Wei Chee
Xu, Hailun
Saha, Sanjay
Fan, Xiaotian
Ong, Hiok Hian
Wang, David Yuchen
Sarkar, Kanchan
Yang, Zhenheng
Guan, Danhui
contents Content moderation remains a critical yet challenging task for large-scale user-generated video platforms, especially in livestreaming environments where moderation must be timely, multimodal, and robust to evolving forms of unwanted content. We present a hybrid moderation framework deployed at production scale that combines supervised classification for known violations with reference-based similarity matching for novel or subtle cases. This hybrid design enables robust detection of both explicit violations and novel edge cases that evade traditional classifiers. Multimodal inputs (text, audio, visual) are processed through both pipelines, with a multimodal large language model (MLLM) distilling knowledge into each to boost accuracy while keeping inference lightweight. In production, the classification pipeline achieves 67% recall at 80% precision, and the similarity pipeline achieves 76% recall at 80% precision. Large-scale A/B tests show a 6-8% reduction in user views of unwanted livestreams}. These results demonstrate a scalable and adaptable approach to multimodal content governance, capable of addressing both explicit violations and emerging adversarial behaviors.
format Preprint
id arxiv_https___arxiv_org_abs_2512_03553
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dynamic Content Moderation in Livestreams: Combining Supervised Classification with MLLM-Boosted Similarity Matching
Yew, Wei Chee
Xu, Hailun
Saha, Sanjay
Fan, Xiaotian
Ong, Hiok Hian
Wang, David Yuchen
Sarkar, Kanchan
Yang, Zhenheng
Guan, Danhui
Computer Vision and Pattern Recognition
Artificial Intelligence
Content moderation remains a critical yet challenging task for large-scale user-generated video platforms, especially in livestreaming environments where moderation must be timely, multimodal, and robust to evolving forms of unwanted content. We present a hybrid moderation framework deployed at production scale that combines supervised classification for known violations with reference-based similarity matching for novel or subtle cases. This hybrid design enables robust detection of both explicit violations and novel edge cases that evade traditional classifiers. Multimodal inputs (text, audio, visual) are processed through both pipelines, with a multimodal large language model (MLLM) distilling knowledge into each to boost accuracy while keeping inference lightweight. In production, the classification pipeline achieves 67% recall at 80% precision, and the similarity pipeline achieves 76% recall at 80% precision. Large-scale A/B tests show a 6-8% reduction in user views of unwanted livestreams}. These results demonstrate a scalable and adaptable approach to multimodal content governance, capable of addressing both explicit violations and emerging adversarial behaviors.
title Dynamic Content Moderation in Livestreams: Combining Supervised Classification with MLLM-Boosted Similarity Matching
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2512.03553