Generative Adversarial Networks and Image Synthesis

Neueste Paper

• Dependency-Aware Parallel Decoding via Attention for Diffusion LLMs

  Bumjun Kim, Dongjae Jeon, Moongyu Jeon, Albert No · 16. März 2026

  Parallel decoding for diffusion LLMs (dLLMs) is difficult because each denoising step provides only token-wise marginal distributions, while unmasking multiple tokens simultaneously requires accounting for inter-token dependencies. We propose Dependency-Aware Parallel Decoding (DAPD), a simple, trai…

• A theory of learning data statistics in diffusion models, from easy to hard

  Lorenzo Bardone, Claudia Merger, Sebastian Goldt · 16. März 2026

  While diffusion models have emerged as a powerful class of generative models, their learning dynamics remain poorly understood. We address this issue first by empirically showing that standard diffusion models trained on natural images exhibit a distributional simplicity bias, learning simple, pair-…

• Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation

  Yichen Zhang, Da Peng, Zonghao Guo, Zijian Zhang, Xuesong Yang, Tong Sun, Shichu Sun, Yidan Zhang, Yanghao Li, Haiyan Zhao, Wang Xu, Qi Shi, Yangang Sun, Chi Chen, Shuo Wang, Yukun Yan, Xu Han, Qiang Ma, Wei Ke, Liang Wang, Zhiyuan Liu, Maosong Sun · 16. März 2026

  A recent cutting-edge topic in multimodal modeling is to unify visual comprehension and generation within a single model. However, the two tasks demand mismatched decoding regimes and visual representations, making it non-trivial to jointly optimize within a shared feature space. In this work, we pr…

• Multiscale Structure-Guided Latent Diffusion for Multimodal MRI Translation

  Jianqiang Lin (Northeastern University, Shenyang, China, Key Laboratory of Intelligent Computing in Medical Image, Shenyang, China), Zhiqiang Shen (Northeastern University, Shenyang, China, Key Laboratory of Intelligent Computing in Medical Image, Shenyang, China), Peng Cao (Northeastern University, Shenyang, China, National Frontiers Science Center for Industrial Intelligence and Systems Optimization, Shenyang, China), Jinzhu Yang (Northeastern University, Shenyang, China, National Frontiers Science Center for Industrial Intelligence and Systems Optimization, Shenyang, China), Osmar R. Zaiane (University of Alberta, Edmonton, Canada), Xiaoli Liu (AiShiWeiLai AI Research, Beijing, China) · 16. März 2026

  Although diffusion models have achieved remarkable progress in multi-modal magnetic resonance imaging (MRI) translation tasks, existing methods still tend to suffer from anatomical inconsistencies or degraded texture details when handling arbitrary missing-modality scenarios. To address these issues…

• Na\"ive PAINE: Lightweight Text-to-Image Generation Improvement with Prompt Evaluation

  Joong Ho Kim, Nicholas Thai, Souhardya Saha Dip, Dong Lao, Keith G. Mills · 16. März 2026

  Text-to-Image (T2I) generation is primarily driven by Diffusion Models (DM) which rely on random Gaussian noise. Thus, like playing the slots at a casino, a DM will produce different results given the same user-defined inputs. This imposes a gambler's burden: To perform multiple generation cycles to…

• Accelerating Diffusion Model Training under Minimal Budgets: A Condensation-Based Perspective

  Rui Huang, Shitong Shao, Zikai Zhou, Pukun Zhao, Hangyu Guo, Tian Ye, Lichen Bai, Shuo Yang, Zeke Xie · 16. März 2026

  Diffusion models have achieved remarkable performance on a wide range of generative tasks, yet training them from scratch is notoriously resource-intensive, typically requiring millions of training images and many GPU days. Motivated by a data-centric view of this bottleneck, we adopt a condensation…

• Curriculum Sampling: A Two-Phase Curriculum for Efficient Training of Flow Matching

  Pengwei Sun · 16. März 2026

  Timestep sampling $p(t)$ is a central design choice in Flow Matching models, yet common practice increasingly favors static middle-biased distributions (e.g., Logit-Normal). We show that this choice induces a speed--quality trade-off: middle-biased sampling accelerates early convergence but yields w…

• Finite Difference Flow Optimization for RL Post-Training of Text-to-Image Models

  David McAllister, Miika Aittala, Tero Karras, Janne Hellsten, Angjoo Kanazawa, Timo Aila, Samuli Laine · 16. März 2026

  Reinforcement learning (RL) has become a standard technique for post-training diffusion-based image synthesis models, as it enables learning from reward signals to explicitly improve desirable aspects such as image quality and prompt alignment. In this paper, we propose an online RL variant that red…

• OpenVision 3: A Family of Unified Visual Encoder for Both Understanding and Generation

  Letian Zhang, Sucheng Ren, Yanqing Liu, Xianhang Li, Zeyu Wang, Yuyin Zhou, Huaxiu Yao, Zeyu Zheng, Weili Nie, Guilin Liu, Zhiding Yu, Cihang Xie · 16. März 2026

  This paper presents a family of advanced vision encoder, named OpenVision 3, that learns a single, unified visual representation that can serve both image understanding and image generation. Our core architecture is simple: we feed VAE-compressed image latents to a ViT encoder and train its output t…

• Stake the Points: Structure-Faithful Instance Unlearning

  Kiseong Hong, JungKyoo Shin, Eunwoo Kim · 16. März 2026

  Machine unlearning (MU) addresses privacy risks in pretrained models. The main goal of MU is to remove the influence of designated data while preserving the utility of retained knowledge. Achieving this goal requires preserving semantic relations among retained instances, which existing studies ofte…

• Asymptotic and Finite-Time Guarantees for Langevin-Based Temperature Annealing in InfoNCE

  Faris Chaudhry · 16. März 2026

  The InfoNCE loss in contrastive learning depends critically on a temperature parameter, yet its dynamics under fixed versus annealed schedules remain poorly understood. We provide a theoretical analysis by modeling embedding evolution under Langevin dynamics on a compact Riemannian manifold. Under m…

• Overcoming the Curvature Bottleneck in MeanFlow

  Xinxi Zhang, Shiwei Tan, Quang Nguyen, Quan Dao, Ligong Han, Xiaoxiao He, Tunyu Zhang, Chengzhi Mao, Dimitris Metaxas, Vladimir Pavlovic · 16. März 2026

  MeanFlow offers a promising framework for one-step generative modeling by directly learning a mean-velocity field, bypassing expensive numerical integration. However, we find that the highly curved generative trajectories of existing models induce a noisy loss landscape, severely bottlenecking conve…

• Disentangled Representation Learning through Unsupervised Symmetry Group Discovery

  Dang-Nhu Barth\'el\'emy, Annabi Louis, Argentieri Sylvain · 13. März 2026

  Symmetry-based disentangled representation learning leverages the group structure of environment transformations to uncover the latent factors of variation. Prior approaches to symmetry-based disentanglement have required strong prior knowledge of the symmetry group's structure, or restrictive assum…

• Diffusion Blend: Inference-Time Multi-Preference Alignment for Diffusion Models

  Min Cheng, Fatemeh Doudi, Dileep Kalathil, Mohammad Ghavamzadeh, Panganamala R. Kumar · 13. März 2026

  Reinforcement learning (RL) algorithms have been used recently to align diffusion models with downstream objectives such as aesthetic quality and text-image consistency by fine-tuning them to maximize a single reward function under a fixed KL regularization. However, this approach is inherently rest…

• GlyphBanana: Advancing Precise Text Rendering Through Agentic Workflows

  Zexuan Yan, Jiarui Jin, Yue Ma, Shijian Wang, Jiahui Hu, Wenxiang Jiao, Yuan Lu, Linfeng Zhang · 13. März 2026

  Despite recent advances in generative models driving significant progress in text rendering, accurately generating complex text and mathematical formulas remains a formidable challenge. This difficulty primarily stems from the limited instruction-following capabilities of current models when encount…

• A Quantitative Characterization of Forgetting in Post-Training

  Krishnakumar Balasubramanian, Shiva Prasad Kasiviswanathan · 13. März 2026

  Continual post-training of generative models is widely used, yet a principled understanding of when and why forgetting occurs remains limited. We develop theoretical results under a two-mode mixture abstraction (representing old and new tasks), proposed by Chen et al. (2025) (arXiv:2510.18874), and …

• A Unified Latent Space Disentanglement VAE Framework with Robust Disentanglement Effectiveness Evaluation

  Xiaoan Lang, Fang Liu · 13. März 2026

  Evaluating and interpreting latent representations, such as variational autoencoders (VAEs), remains a significant challenge for diverse data types, especially when ground-truth generative factors are unknown. To address this, we propose a general framework -- bfVAE -- that unifies several state-of-…

• Bridging Discrete Marks and Continuous Dynamics: Dual-Path Cross-Interaction for Marked Temporal Point Processes

  Yuxiang Liu, Qiao Liu, Tong Luo, Yanglei Gan, Peng He, Yao LIu · 13. März 2026

  Predicting irregularly spaced event sequences with discrete marks poses significant challenges due to the complex, asynchronous dependencies embedded within continuous-time data streams.Existing sequential approaches capture dependencies among event tokens but ignore the continuous evolution between…

• BiGain: Unified Token Compression for Joint Generation and Classification

  Jiacheng Liu, Shengkun Tang, Jiacheng Cui, Dongkuan Xu, Zhiqiang Shen · 13. März 2026

  Acceleration methods for diffusion models (e.g., token merging or downsampling) typically optimize synthesis quality under reduced compute, yet often ignore discriminative capacity. We revisit token compression with a joint objective and present BiGain, a training-free, plug-and-play framework that …

• Coarse-Guided Visual Generation via Weighted h-Transform Sampling

  Yanghao Wang, Ziqi Jiang, Zhen Wang, Long Chen · 13. März 2026

  Coarse-guided visual generation, which synthesizes fine visual samples from degraded or low-fidelity coarse references, is essential for various real-world applications. While training-based approaches are effective, they are inherently limited by high training costs and restricted generalization du…

• The Latent Color Subspace: Emergent Order in High-Dimensional Chaos

  Mateusz Pach, Jessica Bader, Quentin Bouniot, Serge Belongie, Zeynep Akata · 13. März 2026

  Text-to-image generation models have advanced rapidly, yet achieving fine-grained control over generated images remains difficult, largely due to limited understanding of how semantic information is encoded. We develop an interpretation of the color representation in the Variational Autoencoder late…

• Riemannian MeanFlow for One-Step Generation on Manifolds

  Zichen Zhong, Haoliang Sun, Yukun Zhao, Yongshun Gong, Yilong Yin · 12. März 2026

  Flow Matching enables simulation-free training of generative models on Riemannian manifolds, yet sampling typically still relies on numerically integrating a probability-flow ODE. We propose Riemannian MeanFlow (RMF), extending MeanFlow to manifold-valued generation where velocities lie in location-…

• World Models That Know When They Don't Know - Controllable Video Generation with Calibrated Uncertainty

  Zhiting Mei, Tenny Yin, Micah Baker, Ola Shorinwa, Anirudha Majumdar · 12. März 2026

  Recent advances in generative video models have led to significant breakthroughs in high-fidelity video synthesis, specifically in controllable video generation where the generated video is conditioned on text and action inputs, e.g., in instruction-guided video editing and world modeling in robotic…

• Locality-aware Parallel Decoding for Efficient Autoregressive Image Generation

  Zhuoyang Zhang, Luke J. Huang, Chengyue Wu, Shang Yang, Kelly Peng, Yao Lu, Song Han · 12. März 2026

  We present Locality-aware Parallel Decoding (LPD) to accelerate autoregressive image generation. Traditional autoregressive image generation relies on next-patch prediction, a memory-bound process that leads to high latency. Existing works have tried to parallelize next-patch prediction by shifting …

• On the Reliability of Cue Conflict and Beyond

  Pum Jun Kim, Seung-Ah Lee, Seongho Park, Dongyoon Han, Jaejun Yoo · 12. März 2026

  Understanding how neural networks rely on visual cues offers a human-interpretable view of their internal decision processes. The cue-conflict benchmark has been influential in probing shape-texture preference and in motivating the insight that stronger, human-like shape bias is often associated wit…