6 papers
Mechanisms of Width Scaling in Normalized Residual Networks: The Effective Alignment Dimension
Jinhao Zhang, Zeyu Liu, Zicheng Yan +4
Existing theories of neural-network width characterize asymptotic limits, but provide limited guidance on whether an expansion direction identified from finite training data remain…
InsightTok: Improving Text and Face Fidelity in Discrete Tokenization for Autoregressive Image Generation
Yang Yue, Fangyun Wei, Tianyu He +10
Text and faces are among the most perceptually salient and practically important patterns in visual generation, yet they remain challenging for autoregressive generators built on d…
A Qualitative Test-Risk Mechanism for Scaling Behavior in Normalized Residual Networks
Daning Cheng, Zeyu Liu, Jun Sun +4
The scaling behavior, in which test performance often improves as model size and data increase, is a central empirical phenomenon in modern deep learning, yet its theoretical basis…
Refinement via Regeneration: Enlarging Modification Space Boosts Image Refinement in Unified Multimodal Models
Jiayi Guo, Linqing Wang, Jiangshan Wang +6
Unified multimodal models (UMMs) integrate visual understanding and generation within a single framework. For text-to-image (T2I) tasks, this unified capability allows UMMs to refi…
Co-GRPO: Co-Optimized Group Relative Policy Optimization for Masked Diffusion Model
Renping Zhou, Zanlin Ni, Tianyi Chen +6
Recently, Masked Diffusion Models (MDMs) have shown promising potential across vision, language, and cross-modal generation. However, a notable discrepancy exists between their tra…
CODA: Repurposing Continuous VAEs for Discrete Tokenization
Zeyu Liu, Zanlin Ni, Yeguo Hua +4
Discrete visual tokenizers transform images into a sequence of tokens, enabling token-based visual generation akin to language models. However, this process is inherently challengi…