arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2510.04576cs.LGcs.AIcs.CVstat.ML
收藏

SONA:学习条件式、非条件式与感知失配的判别器

SONA: Learning Conditional, Unconditional, and Mismatching-Aware Discriminator

Sony AI(索尼人工智能);Sony Group Corporation(索尼集团);The University of Tokyo(东京大学)

Yuhta Takida, Satoshi Hayakawa, Takashi Shibuya, Masaaki Imaizumi, Naoki Murata, Bac Nguyen, Toshimitsu Uesaka, Chieh-Hsin Lai, Yuki Mitsufuji

AI总结:

针对条件生成对抗网络判别器难以平衡真实性评估与条件对齐的问题,提出整合非条件判别、匹配感知监督与自适应加权的SONA判别器,在多任务上实现更优效果与鲁棒性。

摘要:

深度生成模型在生成复杂内容方面已取得显著进展,但条件生成仍是一项核心挑战。现有条件生成对抗网络的条件判别器往往难以在评估输入样本的真实性与条件对齐性这双重目标间取得平衡。为解决该问题,我们提出一种新型判别器设计,整合三大核心功能:非条件判别、增强对齐敏感度的匹配感知监督,以及动态平衡所有目标的自适应加权机制。具体而言,我们提出了Sum of Naturalness and Alignment(自然度与对齐度之和,SONA),其在最后一层针对自然度(真实性)与对齐度采用带有归纳偏置的独立投影,并配有专用目标函数与自适应加权机制。在类别条件生成任务上的大量实验表明,SONA相较于当前最优方法实现了更优的样本质量与条件对齐效果。此外,我们验证了其在文本生成图像任务中的有效性,证实了所提方法的通用性与鲁棒性。

英文摘要:

Deep generative models have made significant advances in generating complex content, yet conditional generation remains a fundamental challenge. Existing conditional generative adversarial networks often struggle to balance the dual objectives of assessing authenticity and conditional alignment of input samples within their conditional discriminators. To address this, we propose a novel discriminator design that integrates three key capabilities: unconditional discrimination, matching-aware supervision to enhance alignment sensitivity, and adaptive weighting to dynamically balance all objectives. Specifically, we introduce Sum of Naturalness and Alignment (SONA), which employs separate projections for naturalness (authenticity) and alignment in the final layer with an inductive bias, supported by dedicated objective functions and an adaptive weighting mechanism. Extensive experiments on class-conditional generation tasks show that \ours achieves superior sample quality and conditional alignment compared to state-of-the-art methods. Furthermore, we demonstrate its effectiveness in text-to-image generation, confirming the versatility and robustness of our approach.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑