Diffusion models have emerged as a powerful class of generative models across various modalities, including image, video, and audio synthesis. However, their deployment is often limited by significant inference latency, primarily due to the inherently sequential nature of the denoising process. While existing parallelization strategies attempt to accelerate inference by distributing computation across multiple devices, they typically incur high communication overhead, hindering deployment on commercial hardware. To address this challenge, we propose ParaStep, a novel parallelization method based on a reuse-then-predict mechanism that parallelizes diffusion inference by exploiting similarity between adjacent denoising steps. Unlike prior approaches that rely on layer-wise or stage-wise communication, ParaStep employs lightweight, step-wise communication, substantially reducing overhead. ParaStep achieves end-to-end speedups of up to 3.88 on SVD, 2.43 on CogVideoX-2b, and 6.56 on AudioLDM2-large, while maintaining generation quality.

Diffusion models have emerged as a powerful class of generative models across various modalities, including image, video, and audio synthesis. However, their deployment is often limited by significant inference latency, primarily due to the inherently sequential nature of the denoising process. While existing parallelization strategies attempt to accelerate inference by distributing computation across multiple devices, they typically incur high communication overhead, hindering deployment on commercial hardware. To address this challenge, we propose \textbf{ParaStep}, a novel parallelization method based on a reuse-then-predict mechanism that parallelizes diffusion inference by exploiting similarity between adjacent denoising steps. Unlike prior approaches that rely on layer-wise or stage-wise communication, ParaStep employs lightweight, step-wise communication, substantially reducing overhead. ParaStep achieves end-to-end speedups of up to \textbf{3.88}$\times$ on SVD, \textbf{2.43}$\times$ on CogVideoX-2b, and \textbf{6.56}$\times$ on AudioLDM2-large, while maintaining generation quality. These results highlight ParaStep as a scalable and communication-efficient solution for accelerating diffusion inference, particularly in bandwidth-constrained environments.

We are addressing only the major comments in this document. In this document, RXCY refers to Comment Y by Reviewer X. We will ensure to make this crystal clear. In contrast, [4] is an end-to-end reinforcement learning architecture and thus time-consuming. The slowness of CELF in IM is also reported in [2].

In denoising diffusion probabilistic models (DDPMs), the learned noise predictor $ ε_θ ( {\bf x}_t , t)$ is trained to approximate the forward-process noise $ε_t$. The equality $\nabla_{{\bf x}_t} \log q({\bf x}_t) = -\frac 1 {\sqrt {1- {\bar α}_t} } ε_θ ( {\bf x}_t , t)$ plays a fundamental role in both theoretical analyses and algorithmic design, and thus is frequently employed across diffusion-based generative models. In this paper, an explicit formulation of $ ε_θ ( {\bf x}_t , t)$ in terms of the forward-process noise $ε_t$ is derived. This result show how the forward-process noise $ε_t$ contributes to the learned predictor $ ε_θ ( {\bf x}_t , t)$. Furthermore, based on this formulation, we present a novel and mathematically rigorous proof of the fundamental equality above, clarifying its origin and providing new theoretical insight into the structure of diffusion models.

Securing Internet of Things (IoT) devices presents increasing challenges due to their limited computational and energy resources. Radio Frequency Fingerprint Identification (RFFI) emerges as a promising authentication technique to identify wireless devices through hardware impairments. RFFI performance under low signal-to-noise ratio (SNR) scenarios is significantly degraded because the minute hardware features can be easily swamped in noise. In this paper, we leveraged the diffusion model to effectively restore the RFF under low SNR scenarios. Specifically, we trained a powerful noise predictor and tailored a noise removal algorithm to effectively reduce the noise level in the received signal and restore the device fingerprints. We used Wi-Fi as a case study and created a testbed involving 6 commercial off-the-shelf Wi-Fi dongles and a USRP N210 software-defined radio (SDR) platform. We conducted experimental evaluations on various SNR scenarios. The experimental results show that the proposed algorithm can improve the classification accuracy by up to 34.9%.

Guidance is an error-correcting technique used to improve the perceptual quality of images generated by diffusion models. Typically, the correction is achieved by linear extrapolation, using an auxiliary diffusion model that has lower performance than the primary model. Using a 2D toy example, we show that it is highly beneficial when the auxiliary model exhibits similar errors as the primary one but stronger. We verify this finding in higher dimensions, where we show that competitive generative performance to state-of-the-art guidance methods can be achieved when the auxiliary model differs from the primary one only by having stronger weight regularization. As an independent contribution, we investigate whether upweighting long-range spatial dependencies improves visual fidelity. The result is a novel guidance method, which we call sliding window guidance (SWG), that guides the primary model with itself by constraining its receptive field. Intriguingly, SWG aligns better with human preferences than state-of-the-art guidance methods while requiring neither training, architectural modifications, nor class conditioning. The code will be released.

Diffusion models have demonstrated superior performance across various generative tasks including images, videos, and audio. However, they encounter difficulties in directly generating high-resolution samples. Previously proposed solutions to this issue involve modifying the architecture, further training, or partitioning the sampling process into multiple stages. These methods have the limitation of not being able to directly utilize pre-trained models as-is, requiring additional work. In this paper, we introduce upsample guidance, a technique that adapts pretrained diffusion model (e.g., $512^2$) to generate higher-resolution images (e.g., $1536^2$) by adding only a single term in the sampling process. Remarkably, this technique does not necessitate any additional training or relying on external models. We demonstrate that upsample guidance can be applied to various models, such as pixel-space, latent space, and video diffusion models. We also observed that the proper selection of guidance scale can improve image quality, fidelity, and prompt alignment.

Diffusion models generate high-resolution images through iterative stochastic processes. In particular, the denoising method is one of the most popular approaches that predicts the noise in samples and denoises it at each time step. It has been commonly observed that the resolution of generated samples changes over time, starting off blurry and coarse, and becoming sharper and finer. In this paper, we introduce "resolution chromatography" that indicates the signal generation rate of each resolution, which is very helpful concept to mathematically explain this coarse-to-fine behavior in generation process, to understand the role of noise schedule, and to design time-dependent modulation. Using resolution chromatography, we determine which resolution level becomes dominant at a specific time step, and experimentally verify our theory with text-to-image diffusion models. We also propose some direct applications utilizing the concept: upscaling pre-trained models to higher resolutions and time-dependent prompt composing. Our theory not only enables a better understanding of numerous pre-existing techniques for manipulating image generation, but also suggests the potential for designing better noise schedules.

In this work, we study the task of ``visually" translating scene text from a source language (e.g., English) to a target language (e.g., Chinese). Visual translation involves not just the recognition and translation of scene text but also the generation of the translated image that preserves visual features of the text, such as font, size, and background. There are several challenges associated with this task, such as interpolating font to unseen characters and preserving text size and the background. To address these, we introduce VTNet, a novel conditional diffusion-based method. To train the VTNet, we create a synthetic cross-lingual dataset of 600K samples of scene text images in six popular languages, including English, Hindi, Tamil, Chinese, Bengali, and German. We evaluate the performance of VTnet through extensive experiments and comparisons to related methods. Our model also surpasses the previous state-of-the-art results on the conventional scene-text editing benchmarks. Further, we present rigorous qualitative studies to understand the strengths and shortcomings of our model. Results show that our approach generalizes well to unseen words and fonts. We firmly believe our work can benefit real-world applications, such as text translation using a phone camera and translating educational materials. Code and data will be made publicly available.

AI image generation is the most recent AI capability blowing people’s minds (mine included). The ability to create striking visuals from text descriptions has a magical quality to it and points clearly to a shift in how humans create art. The release of Stable Diffusion is a clear milestone in this development because it made a high-performance model available to the masses (performance in terms of image quality, as well as speed and relatively low resource/memory requirements). After experimenting with AI image generation, you may start to wonder how it works. This is a gentle introduction to how Stable Diffusion works. Stable Diffusion is versatile in that it can be used in a number of different ways. Let’s focus at first on image generation from text only (text2img). The image above shows an example text input and the resulting generated image (The actual complete prompt is here). Aside from text to image, another main way of using it is by making it alter images (so inputs are text + image).