Highly Efficient Test-Time Scaling for T2I Diffusion Models with Text Embedding Perturbation

• URL: http://arxiv.org/abs/2512.03996v1
• Date: Wed, 03 Dec 2025 17:27:53 GMT
• Title: Highly Efficient Test-Time Scaling for T2I Diffusion Models with Text Embedding Perturbation
• Authors: Hang Xu, Linjiang Huang, Feng Zhao,
• Abstract summary: Test-time scaling (TTS) aims to achieve better results by increasing random sampling and evaluating samples based on rules and metrics.<n>In this work, we analyze the effects of randomness in T2I diffusion models and explore a new format of randomness for TTS: text embedding perturbation.
• Score: 26.72622200307507
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Test-time scaling (TTS) aims to achieve better results by increasing random sampling and evaluating samples based on rules and metrics. However, in text-to-image(T2I) diffusion models, most related works focus on search strategies and reward models, yet the impact of the stochastic characteristic of noise in T2I diffusion models on the method's performance remains unexplored. In this work, we analyze the effects of randomness in T2I diffusion models and explore a new format of randomness for TTS: text embedding perturbation, which couples with existing randomness like SDE-injected noise to enhance generative diversity and quality. We start with a frequency-domain analysis of these formats of randomness and their impact on generation, and find that these two randomness exhibit complementary behavior in the frequency domain: spatial noise favors low-frequency components (early steps), while text embedding perturbation enhances high-frequency details (later steps), thereby compensating for the potential limitations of spatial noise randomness in high-frequency manipulation. Concurrently, text embedding demonstrates varying levels of tolerance to perturbation across different dimensions of the generation process. Specifically, our method consists of two key designs: (1) Introducing step-based text embedding perturbation, combining frequency-guided noise schedules with spatial noise perturbation. (2) Adapting the perturbation intensity selectively based on their frequency-specific contributions to generation and tolerance to perturbation. Our approach can be seamlessly integrated into existing TTS methods and demonstrates significant improvements on multiple benchmarks with almost no additional computation. Code is available at \href{https://github.com/xuhang07/TEP-Diffusion}{https://github.com/xuhang07/TEP-Diffusion}.

Related papers

• Test-time Adaptive Hierarchical Co-enhanced Denoising Network for Reliable Multimodal Classification [55.56234913868664]
  We propose Test-time Adaptive Hierarchical Co-enhanced Denoising Network (TAHCD) for reliable learning on multimodal data.<n>The proposed method achieves superior classification performance, robustness, and generalization compared with state-of-the-art reliable multimodal learning approaches.
  arXiv  Detail & Related papers  (2026-01-12T03:14:12Z)
• Beyond Randomness: Understand the Order of the Noise in Diffusion [60.3872274940353]
  In text-driven content generation (T2C) diffusion model, semantic of generated content is mostly attributed to the process of text embedding and attention mechanism interaction.<n>This paper conducts a comprehensive analysis of the impact of random noise on the model's generation.<n>We propose a simple but efficient training-free and universal two-step "Semantic Erasure-Injection" process to modulate the initial noise in T2C diffusion model.
  arXiv  Detail & Related papers  (2025-11-11T02:12:38Z)
• Noise Conditional Variational Score Distillation [60.38982038894823]
  Noise Conditional Variational Score Distillation (NCVSD) is a novel method for distilling pretrained diffusion models into generative denoisers.<n>By integrating this insight into the Variational Score Distillation framework, we enable scalable learning of generative denoisers.
  arXiv  Detail & Related papers  (2025-06-11T06:01:39Z)
• Test-Time Scaling of Diffusion Models via Noise Trajectory Search [10.8507840358202]
  We introduce an $epsilon$-greedy search algorithm that globally explores at extreme timesteps and locally exploits during the intermediate steps where de-mixing occurs.<n>Experiments on EDM and Stable Diffusion reveal state-of-the-art scores for class-conditioned/text-to-image generation.
  arXiv  Detail & Related papers  (2025-05-24T19:13:29Z)
• Meta-DiffuB: A Contextualized Sequence-to-Sequence Text Diffusion Model with Meta-Exploration [53.63593099509471]
  We propose a scheduler-exploiter S2S-Diffusion paradigm designed to overcome the limitations of existing S2S-Diffusion models. We employ Meta-Exploration to train an additional scheduler model dedicated to scheduling contextualized noise for each sentence. Our exploiter model, an S2S-Diffusion model, leverages the noise scheduled by our scheduler model for updating and generation.
  arXiv  Detail & Related papers  (2024-10-17T04:06:02Z)
• High-Fidelity Speech Synthesis with Minimal Supervision: All Using Diffusion Models [56.00939852727501]
  Minimally-supervised speech synthesis decouples TTS by combining two types of discrete speech representations. Non-autoregressive framework enhances controllability, and duration diffusion model enables diversified prosodic expression.
  arXiv  Detail & Related papers  (2023-09-27T09:27:03Z)
• DiffSED: Sound Event Detection with Denoising Diffusion [70.18051526555512]
  We reformulate the SED problem by taking a generative learning perspective. Specifically, we aim to generate sound temporal boundaries from noisy proposals in a denoising diffusion process. During training, our model learns to reverse the noising process by converting noisy latent queries to the groundtruth versions.
  arXiv  Detail & Related papers  (2023-08-14T17:29:41Z)
• Adversarial Training of Denoising Diffusion Model Using Dual Discriminators for High-Fidelity Multi-Speaker TTS [0.0]
  The diffusion model is capable of generating high-quality data through a probabilistic approach. It suffers from the drawback of slow generation speed due to the requirement of a large number of time steps. We propose a speech synthesis model with two discriminators: a diffusion discriminator for learning the distribution of the reverse process and a spectrogram discriminator for learning the distribution of the generated data.
  arXiv  Detail & Related papers  (2023-08-03T07:22:04Z)
• Degradation-Noise-Aware Deep Unfolding Transformer for Hyperspectral Image Denoising [9.119226249676501]
  Hyperspectral images (HSIs) are often quite noisy because of narrow band spectral filtering. To reduce the noise in HSI data cubes, both model-driven and learning-based denoising algorithms have been proposed. This paper proposes a Degradation-Noise-Aware Unfolding Network (DNA-Net) that addresses these issues.
  arXiv  Detail & Related papers  (2023-05-06T13:28:20Z)
• Empowering Diffusion Models on the Embedding Space for Text Generation [38.664533078347304]
  We study the optimization challenges encountered with both the embedding space and the denoising model. Data distribution is learnable for embeddings, which may lead to the collapse of the embedding space and unstable training. Based on the above analysis, we propose Difformer, an embedding diffusion model based on Transformer.
  arXiv  Detail & Related papers  (2022-12-19T12:44:25Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.