Diffusion Models

This paper introduces GEAR, a method for training a vector-quantized tokenizer and an autoregressive generator jointly and end-to-end, resolving the issue of non-differentiable VQ indices.

The paper proposes VRPO, a reinforcement learning-based optimization strategy that replaces static alignment losses in diffusion models, significantly improving both convergence and image fidelity.

The paper introduces a histogram-regularized latent diffusion model to synthesize highly realistic and subtype-specific pulmonary nodules in 3D CT volumes, addressing the limitations of existing metho…

The paper proposes FLAME, a novel framework that detects AI-generated image forgeries by identifying intrinsic energy anomalies caused by the diffusion process, achieving state-of-the-art localization…

The paper proposes using a mask-conditioned latent diffusion model to generate synthetic, labeled TEM images for data augmentation, achieving small but measurable performance improvements in defect de…

The paper analyzes that while multimodal large language models (MLLMs) offer superior semantic understanding for image generation, this enhanced capability significantly increases safety risks, partic…

The paper introduces Score-Guided Classification (SGC), a novel framework that uses an unsupervised anomaly score as a 'Pathological Prior' to guide EEG-based depression detection, overcoming the limi…

Dual-Guard introduces a dual-channel latent watermarking framework that simultaneously embeds global provenance and localized content anchors into diffusion images, achieving robust detection against…

The paper introduces SHIFT, a training-free attack that exploits the vulnerability of diffusion-based watermarking by stochastically deflecting the generative trajectory, achieving high removal rates…

This paper proposes ReChannel, a method for dense prediction using a pretrained DiT model, which keeps the encoder but removes the decoder and adapts it with task LoRA. ReChannel maps each token to it…

This paper proposes Twins, a unified continuous token space for multimodal models using ViT and VAE features, and addresses optimization imbalance with a focal regression objective.

The paper introduces Label-Augmented Conditional Diffusion (LACD), a method for learning complex conditional distributions using unlabeled data, and provides theoretical guarantees for its effectivene…

This paper introduces a new approach for high-dimensional one-step generation using a Three-Body Scattering Model (TBSM) with a proper distributional energy.

This paper introduces an attack, PRIVX, demonstrating that even differentially private (DP) Graph Neural Network (GNN) explanations leak enough structural information to allow an adversary to accurate…

The Parameterized Diffusion Policy (PDP) framework transforms diffusion models from general stochastic generators into precise, steerable tools for learning and adapting complex robotic behaviors by e…

The paper proposes a fast and lightweight novel view synthesis method using a differentiable Multiplane Image (MPI) representation, achieving significant speed and size improvements over state-of-the-…

The paper introduces diffGHOST, a conditional diffusion model that generates synthetic, privacy-preserving mobility trajectories by explicitly mitigating sample memorization in the latent space.

DiffCrossGait proposes a novel trajectory-level alignment method using latent diffusion to overcome domain discrepancies in 2D-3D gait recognition, achieving state-of-the-art performance.

The paper repurposes a pre-trained speech classifier as the backbone for diffusion generation, reducing the need for two separately trained models.

The paper proposes a visual program synthesis framework using a VLM to generate accurate training data for semiconductor inspection, mitigating the sim-to-real gap by applying input binarization to st…