Aesthetic Alignment Risks Assimilation: How Image Generation and Reward Models Reinforce Beauty Bias and Ideological "Censorship"

• URL: http://arxiv.org/abs/2512.11883v1
• Date: Tue, 09 Dec 2025 00:24:29 GMT
• Title: Aesthetic Alignment Risks Assimilation: How Image Generation and Reward Models Reinforce Beauty Bias and Ideological "Censorship"
• Authors: Wenqi Marshall Guo, Qingyun Qian, Khalad Hasan, Shan Du,
• Abstract summary: Over-aligning image generation models to a generalized aesthetic preference conflicts with user intent.<n>We test this bias by constructing a wide-spectrum aesthetics dataset and evaluating state-of-the-art generation and reward models.
• Score: 10.879152680774318
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Over-aligning image generation models to a generalized aesthetic preference conflicts with user intent, particularly when ``anti-aesthetic" outputs are requested for artistic or critical purposes. This adherence prioritizes developer-centered values, compromising user autonomy and aesthetic pluralism. We test this bias by constructing a wide-spectrum aesthetics dataset and evaluating state-of-the-art generation and reward models. We find that aesthetic-aligned generation models frequently default to conventionally beautiful outputs, failing to respect instructions for low-quality or negative imagery. Crucially, reward models penalize anti-aesthetic images even when they perfectly match the explicit user prompt. We confirm this systemic bias through image-to-image editing and evaluation against real abstract artworks.

Related papers

• Enhancing Reward Models for High-quality Image Generation: Beyond Text-Image Alignment [63.823383517957986]
  We propose a novel evaluation score, ICT (Image-Contained-Text) score, that achieves and surpasses the objectives of text-image alignment.<n>We further train an HP (High-Preference) score model using solely the image modality to enhance image aesthetics and detail quality.
  arXiv  Detail & Related papers  (2025-07-25T07:01:50Z)
• Aesthetics Without Semantics [3.644950723229025]
  We create a database of images with minimal semantic content and devise a method to generate images on the ugly side of aesthetic valuations.<n>We show how augmenting an image set biased towards beautiful images with ugly images can modify, or even invert, an observed relationship between image features and aesthetics valuation.
  arXiv  Detail & Related papers  (2025-05-08T15:22:11Z)
• Anti-Aesthetics: Protecting Facial Privacy against Customized Text-to-Image Synthesis [18.743360923809217]
  We propose a Hierarchical Anti-Aesthetic (HAA) framework to fully explore aesthetic cues.<n>Our HAA effectively achieves the goal of anti-aesthetics from a global to a local level during customized generation.
  arXiv  Detail & Related papers  (2025-04-16T14:44:00Z)
• Aligning Vision Models with Human Aesthetics in Retrieval: Benchmarks and Algorithms [91.19304518033144]
  We aim to align vision models with human aesthetic standards in a retrieval system. We propose a preference-based reinforcement learning method that fine-tunes the vision models to better align the vision models with human aesthetics.
  arXiv  Detail & Related papers  (2024-06-13T17:59:20Z)
• Inpaint Biases: A Pathway to Accurate and Unbiased Image Generation [0.0]
  We introduce the Inpaint Biases framework, which employs user-defined masks and inpainting techniques to enhance the accuracy of image generation. We demonstrate how this framework significantly improves the fidelity of generated images to the user's intent, thereby expanding the models' creative capabilities.
  arXiv  Detail & Related papers  (2024-05-29T05:04:07Z)
• Human Aesthetic Preference-Based Large Text-to-Image Model Personalization: Kandinsky Generation as an Example [4.744780823386797]
  This paper introduces a prompting-free generative approach that empowers users to automatically generate personalized painterly content. By relying on the user's aesthetic evaluation and preference for the artist model-generated images, this approach creates the user a personalized model.
  arXiv  Detail & Related papers  (2024-02-09T13:11:19Z)
• New Job, New Gender? Measuring the Social Bias in Image Generation Models [85.26441602999014]
  Image generation models are susceptible to generating content that perpetuates social stereotypes and biases. We propose BiasPainter, a framework that can accurately, automatically and comprehensively trigger social bias in image generation models. BiasPainter can achieve 90.8% accuracy on automatic bias detection, which is significantly higher than the results reported in previous work.
  arXiv  Detail & Related papers  (2024-01-01T14:06:55Z)
• Impressions: Understanding Visual Semiotics and Aesthetic Impact [66.40617566253404]
  We present Impressions, a novel dataset through which to investigate the semiotics of images. We show that existing multimodal image captioning and conditional generation models struggle to simulate plausible human responses to images. This dataset significantly improves their ability to model impressions and aesthetic evaluations of images through fine-tuning and few-shot adaptation.
  arXiv  Detail & Related papers  (2023-10-27T04:30:18Z)
• VILA: Learning Image Aesthetics from User Comments with Vision-Language Pretraining [53.470662123170555]
  We propose learning image aesthetics from user comments, and exploring vision-language pretraining methods to learn multimodal aesthetic representations. Specifically, we pretrain an image-text encoder-decoder model with image-comment pairs, using contrastive and generative objectives to learn rich and generic aesthetic semantics without human labels. Our results show that our pretrained aesthetic vision-language model outperforms prior works on image aesthetic captioning over the AVA-Captions dataset.
  arXiv  Detail & Related papers  (2023-03-24T23:57:28Z)
• Modeling, Quantifying, and Predicting Subjectivity of Image Aesthetics [21.46956783120668]
  We propose a novel unified probabilistic framework that can model and quantify subjective aesthetic preference based on the subjective logic. In this framework, the rating distribution is modeled as a beta distribution, from which the probabilities of being definitely pleasing, being definitely unpleasing, and being uncertain can be obtained. We present a method to learn deep neural networks for prediction of image aesthetics, which is shown to be effective in improving the performance of subjectivity prediction via experiments.
  arXiv  Detail & Related papers  (2022-08-20T12:16:45Z)
• User-Guided Personalized Image Aesthetic Assessment based on Deep Reinforcement Learning [64.07820203919283]
  We propose a novel user-guided personalized image aesthetic assessment framework. It leverages user interactions to retouch and rank images for aesthetic assessment based on deep reinforcement learning (DRL) It generates personalized aesthetic distribution that is more in line with the aesthetic preferences of different users.
  arXiv  Detail & Related papers  (2021-06-14T15:19:48Z)

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.