From Sparse Decisions to Dense Reasoning: A Multi-attribute Trajectory Paradigm for Multimodal Moderation

Tianle Gu1,2, Kexin Huang3, Lingyu Li2, Ruilin Luo1, Shiyang Huang2, Zongqi Wang1, Yujiu Yang1, Yan Teng2, Yingchun Wang2,
1Tsinghua University
2Shanghai Artificial Intelligence Laboratory
3Fudan University

Provide an image and an optional text prompt as input. The model generates the corresponding response in the output panel.

This video demonstrates the UniRM model.

Abstract

Safety moderation is pivotal for identifying harmful content. Despite the success of textual safety moderation, its multimodal counterparts remain hindered by a dual sparsity of data and supervision. Conventional reliance on binary labels lead to shortcut learning, which obscures the intrinsic classification boundaries necessary for effective multimodal discrimination. Hence, we propose a novel learning paradigm (UniMod) that transitions from sparse decision-making to dense reasoning traces. By constructing structured trajectories encompassing evidence grounding, modality assessment, risk mapping, policy decision, and response generation, we reformulate monolithic decision tasks into a multi-dimensional boundary learning process. This approach forces the model to ground its decision in explicit safety semantics, preventing the model from converging on superficial shortcuts. To facilitate this paradigm, we develop a multi-head scalar reward model (UniRM). UniRM provides multi-dimensional supervision by assigning attribute-level scores to the response generation stage. Furthermore, we introduce specialized optimization strategies to decouple task-specific parameters and rebalance training dynamics, effectively resolving interference between diverse objectives in multi-task learning. Empirical results show UniMod achieves competitive textual moderation performance and sets a new multimodal state-of-the-art (SOTA) using less than 40% of the training data used by leading baselines. Ablations further validate our multi-attribute trajectory reasoning, offering an effective and efficient framework for multimodal moderation.

Cite Us

@misc{gu2026sparsedecisionsdensereasoning,
      title={From Sparse Decisions to Dense Reasoning: A Multi-attribute Trajectory Paradigm for Multimodal Moderation}, 
      author={Tianle Gu and Kexin Huang and Lingyu Li and Ruilin Luo and Shiyang Huang and Zongqi Wang and Yujiu Yang and Yan Teng and Yingchun Wang},
      year={2026},
      eprint={2602.02536},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2602.02536}, 
}