PopAlign: Population-Level Alignment for Fair Text-to-Image Generation

• URL: http://arxiv.org/abs/2406.19668v1
• Date: Fri, 28 Jun 2024 05:38:32 GMT
• Title: PopAlign: Population-Level Alignment for Fair Text-to-Image Generation
• Authors: Shufan Li, Harkanwar Singh, Aditya Grover,
• Abstract summary: We introduce PopAlign, a novel approach for population-level preference optimization. We show that PopAlign significantly mitigates the bias of pretrained T2I models while largely preserving the generation quality.
• Score: 26.457571615782985
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Text-to-image (T2I) models achieve high-fidelity generation through extensive training on large datasets. However, these models may unintentionally pick up undesirable biases of their training data, such as over-representation of particular identities in gender or ethnicity neutral prompts. Existing alignment methods such as Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) fail to address this problem effectively because they operate on pairwise preferences consisting of individual samples, while the aforementioned biases can only be measured at a population level. For example, a single sample for the prompt "doctor" could be male or female, but a model generating predominantly male doctors even with repeated sampling reflects a gender bias. To address this limitation, we introduce PopAlign, a novel approach for population-level preference optimization, while standard optimization would prefer entire sets of samples over others. We further derive a stochastic lower bound that directly optimizes for individual samples from preferred populations over others for scalable training. Using human evaluation and standard image quality and bias metrics, we show that PopAlign significantly mitigates the bias of pretrained T2I models while largely preserving the generation quality. Code is available at https://github.com/jacklishufan/PopAlignSDXL.

Related papers

• SePPO: Semi-Policy Preference Optimization for Diffusion Alignment [67.8738082040299]
  We propose a preference optimization method that aligns DMs with preferences without relying on reward models or paired human-annotated data. We validate SePPO across both text-to-image and text-to-video benchmarks.
  arXiv  Detail & Related papers  (2024-10-07T17:56:53Z)
• Tackling the Problem of Distributional Shifts: Correcting Misspecified, High-Dimensional Data-Driven Priors for Inverse Problems [39.58317527488534]
  Data-driven population-level distributions are emerging as an appealing alternative to simple parametric priors in inverse problems. It is difficult to acquire independent and identically distributed samples from the underlying data-generating process of interest to train these models. We show that starting from a misspecified prior distribution, the updated distribution becomes progressively closer to the underlying population-level distribution.
  arXiv  Detail & Related papers  (2024-07-24T22:39:27Z)
• Curriculum Direct Preference Optimization for Diffusion and Consistency Models [110.08057135882356]
  We propose a novel and enhanced version of DPO based on curriculum learning for text-to-image generation. Our approach, Curriculum DPO, is compared against state-of-the-art fine-tuning approaches on three benchmarks.
  arXiv  Detail & Related papers  (2024-05-22T13:36:48Z)
• Debiased Model-based Interactive Recommendation [22.007617148466807]
  We develop a model called textbfidentifiable textbfDebiased textbfModel-based textbfInteractive textbfRecommendation (textbfiDMIR in short) For the first drawback, we devise a debiased causal world model based on the causal mechanism of the time-varying recommendation generation process with identification guarantees. For the second drawback, we devise a debiased contrastive policy, which coincides with the debiased contrastive learning and avoids sampling bias
  arXiv  Detail & Related papers  (2024-02-24T14:10:04Z)
• Unbiased Image Synthesis via Manifold Guidance in Diffusion Models [9.531220208352252]
  Diffusion Models often inadvertently favor certain data attributes, undermining the diversity of generated images. We propose a plug-and-play method named Manifold Sampling Guidance, which is also the first unsupervised method to mitigate bias issue in DDPMs.
  arXiv  Detail & Related papers  (2023-07-17T02:03:17Z)
• Language Models Get a Gender Makeover: Mitigating Gender Bias with Few-Shot Data Interventions [50.67412723291881]
  Societal biases present in pre-trained large language models are a critical issue. We propose data intervention strategies as a powerful yet simple technique to reduce gender bias in pre-trained models.
  arXiv  Detail & Related papers  (2023-06-07T16:50:03Z)
• BLIND: Bias Removal With No Demographics [29.16221451643288]
  We introduce BLIND, a method for bias removal with no prior knowledge of the demographics in the dataset. While training a model on a downstream task, BLIND detects biased samples using an auxiliary model that predicts the main model's success, and down-weights those samples during the training process. Experiments with racial and gender biases in sentiment classification and occupation classification tasks demonstrate that BLIND mitigates social biases without relying on a costly demographic annotation process.
  arXiv  Detail & Related papers  (2022-12-20T18:59:42Z)
• Bias Mimicking: A Simple Sampling Approach for Bias Mitigation [57.17709477668213]
  We introduce a new class-conditioned sampling method: Bias Mimicking. Bias Mimicking improves underrepresented groups' accuracy of sampling methods by 3% over four benchmarks.
  arXiv  Detail & Related papers  (2022-09-30T17:33:00Z)
• Improving Gender Fairness of Pre-Trained Language Models without Catastrophic Forgetting [88.83117372793737]
  Forgetting information in the original training data may damage the model's downstream performance by a large margin. We propose GEnder Equality Prompt (GEEP) to improve gender fairness of pre-trained models with less forgetting.
  arXiv  Detail & Related papers  (2021-10-11T15:52:16Z)
• Adversarial Examples Generation for Reducing Implicit Gender Bias in Pre-trained Models [2.6329024988388925]
  We propose a method to automatically generate implicit gender bias samples at sentence-level and a metric to measure gender bias. The metric will be used to guide the generation of examples from Pre-trained models. Therefore, those examples could be used to impose attacks on Pre-trained Models.
  arXiv  Detail & Related papers  (2021-10-03T20:22:54Z)
• Improving Maximum Likelihood Training for Text Generation with Density Ratio Estimation [51.091890311312085]
  We propose a new training scheme for auto-regressive sequence generative models, which is effective and stable when operating at large sample space encountered in text generation. Our method stably outperforms Maximum Likelihood Estimation and other state-of-the-art sequence generative models in terms of both quality and diversity.
  arXiv  Detail & Related papers  (2020-07-12T15:31:24Z)

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.