Generative Adversarial Networks and Image Synthesis

Latest papers

• From Kernels to Attention: A Transformer Framework for Density and Score Estimation

  Vasily Ilin, Peter Sushko · 11 November 2025

  We introduce a unified attention-based framework for joint score and density estimation. Framing the problem as a sequence-to-sequence task, we develop a permutation- and affine-equivariant transformer that estimates both the probability density $f(x)$ and its score $\nabla_x \log f(x)$ directly fro…

• TuckA: Hierarchical Compact Tensor Experts for Efficient Fine-Tuning

  Qifeng Lei, Zhiyong Yang, Qianqian Xu, Cong Hua, Peisong Wen, Qingming Huang · 11 November 2025

  Efficiently fine-tuning pre-trained models for downstream tasks is a key challenge in the era of foundation models. Parameter-efficient fine-tuning (PEFT) presents a promising solution, achieving performance comparable to full fine-tuning by updating only a small number of adaptation weights per lay…

• Inference-Time Scaling of Diffusion Models for Infrared Data Generation

  Kai A. Horstmann, Maxim Clouser, Kia Khezeli · 11 November 2025

  Infrared imagery enables temperature-based scene understanding using passive sensors, particularly under conditions of low visibility where traditional RGB imaging fails. Yet, developing downstream vision models for infrared applications is hindered by the scarcity of high-quality annotated data, du…

• KLASS: KL-Guided Fast Inference in Masked Diffusion Models

  Seo Hyun Kim, Sunwoo Hong, Hojung Jung, Youngrok Park, Se-Young Yun · 11 November 2025

  Masked diffusion models have demonstrated competitive results on various tasks including language generation. However, due to its iterative refinement process, the inference is often bottlenecked by slow and static sampling speed. To overcome this problem, we introduce `KL-Adaptive Stability Samplin…

• Effective Test-Time Scaling of Discrete Diffusion through Iterative Refinement

  Sanghyun Lee, Sunwoo Kim, Seungryong Kim, Jongho Park, Dongmin Park · 11 November 2025

  Test-time scaling through reward-guided generation remains largely unexplored for discrete diffusion models despite its potential as a promising alternative. In this work, we introduce Iterative Reward-Guided Refinement (IterRef), a novel test-time scaling method tailored to discrete diffusion that …

• Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion

  Xun Huang, Zhengqi Li, Guande He, Mingyuan Zhou, Eli Shechtman · 11 November 2025

  We introduce Self Forcing, a novel training paradigm for autoregressive video diffusion models. It addresses the longstanding issue of exposure bias, where models trained on ground-truth context must generate sequences conditioned on their own imperfect outputs during inference. Unlike prior methods…

• Rotary Masked Autoencoders are Versatile Learners

  Uros Zivanovic, Serafina Di Gioia, Andre Scaffidi, Mart\'in de los Rios, Gabriella Contardo, Roberto Trotta · 11 November 2025

  Applying Transformers to irregular time-series typically requires specializations to their baseline architecture, which can result in additional computational overhead and increased method complexity. We present the Rotary Masked Autoencoder (RoMAE), which utilizes the popular Rotary Positional Embe…

• Unveiling the Training Dynamics of ReLU Networks through a Linear Lens

  Longqing Ye · 11 November 2025

  Deep neural networks, particularly those employing Rectified Linear Units (ReLU), are often perceived as complex, high-dimensional, non-linear systems. This complexity poses a significant challenge to understanding their internal learning mechanisms. In this work, we propose a novel analytical frame…

• Visual Exploration of Feature Relationships in Sparse Autoencoders with Curated Concepts

  Xinyuan Yan, Shusen Liu, Kowshik Thopalli, Bei Wang · 11 November 2025

  Sparse autoencoders (SAEs) have emerged as a powerful tool for uncovering interpretable features in large language models (LLMs) through the sparse directions they learn. However, the sheer number of extracted directions makes comprehensive exploration intractable. While conventional embedding techn…

• Depth-induced NTK: Bridging Over-parameterized Neural Networks and Deep Neural Kernels

  Yong-Ming Tian, Shuang Liang, Shao-Qun Zhang, Feng-Lei Fan · 11 November 2025

  While deep learning has achieved remarkable success across a wide range of applications, its theoretical understanding of representation learning remains limited. Deep neural kernels provide a principled framework to interpret over-parameterized neural networks by mapping hierarchical feature transf…

• Understanding Representation Dynamics of Diffusion Models via Low-Dimensional Modeling

  Xiao Li, Zekai Zhang, Xiang Li, Siyi Chen, Zhihui Zhu, Peng Wang, Qing Qu · 11 November 2025

  Diffusion models, though originally designed for generative tasks, have demonstrated impressive self-supervised representation learning capabilities. A particularly intriguing phenomenon in these models is the emergence of unimodal representation dynamics, where the quality of learned features peaks…

• Enhancing Diffusion Model Guidance through Calibration and Regularization

  Seyed Alireza Javid, Amirhossein Bagheri, Nuria Gonz\'alez-Prelcic · 11 November 2025

  Classifier-guided diffusion models have emerged as a powerful approach for conditional image generation, but they suffer from overconfident predictions during early denoising steps, causing the guidance gradient to vanish. This paper introduces two complementary contributions to address this issue. …

• Sublinear iterations can suffice even for DDPMs

  Matthew S. Zhang, Stephen Huan, Jerry Huang, Nicholas M. Boffi, Sitan Chen, Sinho Chewi · 10 November 2025

  SDE-based methods such as denoising diffusion probabilistic models (DDPMs) have shown remarkable success in real-world sample generation tasks. Prior analyses of DDPMs have been focused on the exponential Euler discretization, showing guarantees that generally depend at least linearly on the dimensi…

• CPO: Condition Preference Optimization for Controllable Image Generation

  Zonglin Lyu, Ming Li, Xinxin Liu, Chen Chen · 10 November 2025

  To enhance controllability in text-to-image generation, ControlNet introduces image-based control signals, while ControlNet++ improves pixel-level cycle consistency between generated images and the input control signal. To avoid the prohibitive cost of back-propagating through the sampling process, …

• Learning Fourier shapes to probe the geometric world of deep neural networks

  Jian Wang, Yixing Yong, Haixia Bi, Lijun He, Fan Li · 10 November 2025

  While both shape and texture are fundamental to visual recognition, research on deep neural networks (DNNs) has predominantly focused on the latter, leaving their geometric understanding poorly probed. Here, we show: first, that optimized shapes can act as potent semantic carriers, generating high-c…

• Explicit Density Approximation for Neural Implicit Samplers Using a Bernstein-Based Convex Divergence

  Jos\'e Manuel de Frutos, Manuel A. V\'azquez, Pablo M. Olmos, Joaqu\'in M\'iguez · 7 November 2025

  Rank-based statistical metrics, such as the invariant statistical loss (ISL), have recently emerged as robust and practically effective tools for training implicit generative models. In this work, we introduce dual-ISL, a novel likelihood-free objective for training implicit generative models that i…

• Towards Causal Market Simulators

  Dennis Thumm, Luis Ontaneda Mijares · 7 November 2025

  Market generators using deep generative models have shown promise for synthetic financial data generation, but existing approaches lack causal reasoning capabilities essential for counterfactual analysis and risk assessment. We propose a Time-series Neural Causal Model VAE (TNCM-VAE) that combines v…

• Optimal Inference Schedules for Masked Diffusion Models

  Sitan Chen, Kevin Cong, Jerry Li · 7 November 2025

  A major bottleneck of standard auto-regressive large language models is that their inference process is inherently sequential, resulting in very long and costly inference times. To circumvent this, practitioners proposed a class of language models called diffusion language models, of which the maske…

• Provable Separations between Memorization and Generalization in Diffusion Models

  Zeqi Ye, Qijie Zhu, Molei Tao, Minshuo Chen · 6 November 2025

  Diffusion models have achieved remarkable success across diverse domains, but they remain vulnerable to memorization -- reproducing training data rather than generating novel outputs. This not only limits their creative potential but also raises concerns about privacy and safety. While empirical stu…

• When Generative Artificial Intelligence meets Extended Reality: A Systematic Review

  Xinyu Ning, Yan Zhuo, Xian Wang, Chan-In Devin Sio, Lik-Hang Lee · 6 November 2025

  With the continuous advancement of technology, the application of generative artificial intelligence (AI) in various fields is gradually demonstrating great potential, particularly when combined with Extended Reality (XR), creating unprecedented possibilities. This survey article systematically revi…

• ForTIFAI: Fending Off Recursive Training Induced Failure for AI Model Collapse

  Soheil Zibakhsh Shabgahi, Pedram Aghazadeh, Azalia Mirhoseini, Farinaz Koushanfar · 6 November 2025

  The increasing reliance on generative AI models is rapidly increasing the volume of synthetic data, with some projections suggesting that most available new data for training could be machine-generated by 2030. This shift to a mainly synthetic content presents a critical challenge: repeated training…

• Why Less is More (Sometimes): A Theory of Data Curation

  Elvis Dohmatob, Mohammad Pezeshki, Reyhane Askari-Hemmat · 6 November 2025

  This paper introduces a theoretical framework to resolve a central paradox in modern machine learning: When is it better to use less data? This question has become critical as classical scaling laws suggesting ``more is more'' (Sun et al., 2025) are challenged by methods like LIMO (``less is more'')…

• TAUE: Training-free Noise Transplant and Cultivation Diffusion Model

  Daichi Nagai, Ryugo Morita, Shunsuke Kitada, Hitoshi Iyatomi · 5 November 2025

  Despite the remarkable success of text-to-image diffusion models, their output of a single, flattened image remains a critical bottleneck for professional applications requiring layer-wise control. Existing solutions either rely on fine-tuning with large, inaccessible datasets or are training-free y…

• Autoencoding Random Forests

  Binh Duc Vu, Jan Kapar, Marvin Wright, David S. Watson · 5 November 2025

  We propose a principled method for autoencoding with random forests. Our strategy builds on foundational results from nonparametric statistics and spectral graph theory to learn a low-dimensional embedding of the model that optimally represents relationships in the data. We provide exact and approxi…

• MediQ-GAN: Quantum-Inspired GAN for High Resolution Medical Image Generation

  Qingyue Jiao, Yongcan Tang, Jun Zhuang, Jason Cong, Yiyu Shi · 5 November 2025

  Machine learning-assisted diagnosis shows promise, yet medical imaging datasets are often scarce, imbalanced, and constrained by privacy, making data augmentation essential. Classical generative models typically demand extensive computational and sample resources. Quantum computing offers a promisin…