Understanding Implosion in Text-to-Image Generative Models

• URL: http://arxiv.org/abs/2409.12314v1
• Date: Wed, 18 Sep 2024 21:06:45 GMT
• Title: Understanding Implosion in Text-to-Image Generative Models
• Abstract summary: We establish the first analytical framework on robustness of image generative models to poisoning attacks. We quantify the impact of training data by the hardness of alignment, measured by an Alignment Difficulty metric. We prove that AD increases with the number of individual prompts (or concepts) poisoned. As a result, the generative model implodes and outputs random, incoherent images at large.
• Score: 26.90376728384208
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Recent works show that text-to-image generative models are surprisingly vulnerable to a variety of poisoning attacks. Empirical results find that these models can be corrupted by altering associations between individual text prompts and associated visual features. Furthermore, a number of concurrent poisoning attacks can induce "model implosion," where the model becomes unable to produce meaningful images for unpoisoned prompts. These intriguing findings highlight the absence of an intuitive framework to understand poisoning attacks on these models. In this work, we establish the first analytical framework on robustness of image generative models to poisoning attacks, by modeling and analyzing the behavior of the cross-attention mechanism in latent diffusion models. We model cross-attention training as an abstract problem of "supervised graph alignment" and formally quantify the impact of training data by the hardness of alignment, measured by an Alignment Difficulty (AD) metric. The higher the AD, the harder the alignment. We prove that AD increases with the number of individual prompts (or concepts) poisoned. As AD grows, the alignment task becomes increasingly difficult, yielding highly distorted outcomes that frequently map meaningful text prompts to undefined or meaningless visual representations. As a result, the generative model implodes and outputs random, incoherent images at large. We validate our analytical framework through extensive experiments, and we confirm and explain the unexpected (and unexplained) effect of model implosion while producing new, unforeseen insights. Our work provides a useful tool for studying poisoning attacks against diffusion models and their defenses.

Related papers

• Deferred Poisoning: Making the Model More Vulnerable via Hessian Singularization [39.37308843208039]
  We introduce a more threatening type of poisoning attack called the Deferred Poisoning Attack. This new attack allows the model to function normally during the training and validation phases but makes it very sensitive to evasion attacks or even natural noise. We have conducted both theoretical and empirical analyses of the proposed method and validated its effectiveness through experiments on image classification tasks.
  arXiv  Detail & Related papers  (2024-11-06T08:27:49Z)
• Model X-ray:Detecting Backdoored Models via Decision Boundary [62.675297418960355]
  Backdoor attacks pose a significant security vulnerability for deep neural networks (DNNs) We propose Model X-ray, a novel backdoor detection approach based on the analysis of illustrated two-dimensional (2D) decision boundaries. Our approach includes two strategies focused on the decision areas dominated by clean samples and the concentration of label distribution.
  arXiv  Detail & Related papers  (2024-02-27T12:42:07Z)
• Diffusion Model with Cross Attention as an Inductive Bias for Disentanglement [58.9768112704998]
  Disentangled representation learning strives to extract the intrinsic factors within observed data. We introduce a new perspective and framework, demonstrating that diffusion models with cross-attention can serve as a powerful inductive bias. This is the first work to reveal the potent disentanglement capability of diffusion models with cross-attention, requiring no complex designs.
  arXiv  Detail & Related papers  (2024-02-15T05:07:54Z)
• Adversarial Examples are Misaligned in Diffusion Model Manifolds [7.979892202477701]
  This study is dedicated to the investigation of adversarial attacks through the lens of diffusion models. Our focus lies in utilizing the diffusion model to detect and analyze the anomalies introduced by these attacks on images. Results demonstrate a notable capacity to discriminate effectively between benign and attacked images.
  arXiv  Detail & Related papers  (2024-01-12T15:29:21Z)
• Exploring Model Dynamics for Accumulative Poisoning Discovery [62.08553134316483]
  We propose a novel information measure, namely, Memorization Discrepancy, to explore the defense via the model-level information. By implicitly transferring the changes in the data manipulation to that in the model outputs, Memorization Discrepancy can discover the imperceptible poison samples. We thoroughly explore its properties and propose Discrepancy-aware Sample Correction (DSC) to defend against accumulative poisoning attacks.
  arXiv  Detail & Related papers  (2023-06-06T14:45:24Z)
• Trade-offs in Fine-tuned Diffusion Models Between Accuracy and Interpretability [5.865936619867771]
  We unravel a consequential trade-off between image fidelity as gauged by conventional metrics and model interpretability in generative diffusion models. We present a set of design principles for the development of truly interpretable generative models.
  arXiv  Detail & Related papers  (2023-03-31T09:11:26Z)
• Person Image Synthesis via Denoising Diffusion Model [116.34633988927429]
  We show how denoising diffusion models can be applied for high-fidelity person image synthesis. Our results on two large-scale benchmarks and a user study demonstrate the photorealism of our proposed approach under challenging scenarios.
  arXiv  Detail & Related papers  (2022-11-22T18:59:50Z)
• Data Poisoning Attacks Against Multimodal Encoders [24.02062380303139]
  We study poisoning attacks against multimodal models in both visual and linguistic modalities. To mitigate the attacks, we propose both pre-training and post-training defenses.
  arXiv  Detail & Related papers  (2022-09-30T06:50:08Z)
• Proactive Pseudo-Intervention: Causally Informed Contrastive Learning For Interpretable Vision Models [103.64435911083432]
  We present a novel contrastive learning strategy called it Proactive Pseudo-Intervention (PPI) PPI leverages proactive interventions to guard against image features with no causal relevance. We also devise a novel causally informed salience mapping module to identify key image pixels to intervene, and show it greatly facilitates model interpretability.
  arXiv  Detail & Related papers  (2020-12-06T20:30:26Z)
• A General Framework for Survival Analysis and Multi-State Modelling [70.31153478610229]
  We use neural ordinary differential equations as a flexible and general method for estimating multi-state survival models. We show that our model exhibits state-of-the-art performance on popular survival data sets and demonstrate its efficacy in a multi-state setting.
  arXiv  Detail & Related papers  (2020-06-08T19:24:54Z)

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.