Curriculum Direct Preference Optimization for Diffusion and Consistency Models

• URL: http://arxiv.org/abs/2405.13637v2
• Date: Fri, 24 May 2024 13:14:40 GMT
• Title: Curriculum Direct Preference Optimization for Diffusion and Consistency Models
• Authors: Florinel-Alin Croitoru, Vlad Hondru, Radu Tudor Ionescu, Nicu Sebe, Mubarak Shah,
• Abstract summary: 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.
• Score: 110.08057135882356
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Direct Preference Optimization (DPO) has been proposed as an effective and efficient alternative to reinforcement learning from human feedback (RLHF). In this paper, we propose a novel and enhanced version of DPO based on curriculum learning for text-to-image generation. Our method is divided into two training stages. First, a ranking of the examples generated for each prompt is obtained by employing a reward model. Then, increasingly difficult pairs of examples are sampled and provided to a text-to-image generative (diffusion or consistency) model. Generated samples that are far apart in the ranking are considered to form easy pairs, while those that are close in the ranking form hard pairs. In other words, we use the rank difference between samples as a measure of difficulty. The sampled pairs are split into batches according to their difficulty levels, which are gradually used to train the generative model. Our approach, Curriculum DPO, is compared against state-of-the-art fine-tuning approaches on three benchmarks, outperforming the competing methods in terms of text alignment, aesthetics and human preference. Our code is available at https://anonymous.4open.science/r/Curriculum-DPO-EE14.

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)
• PopAlign: Population-Level Alignment for Fair Text-to-Image Generation [26.457571615782985]
  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.
  arXiv  Detail & Related papers  (2024-06-28T05:38:32Z)
• Pre-Trained Vision-Language Models as Partial Annotators [40.89255396643592]
  Pre-trained vision-language models learn massive data to model unified representations of images and natural languages. In this paper, we investigate a novel "pre-trained annotating - weakly-supervised learning" paradigm for pre-trained model application and experiment on image classification tasks.
  arXiv  Detail & Related papers  (2024-05-23T17:17:27Z)
• Diversified in-domain synthesis with efficient fine-tuning for few-shot classification [64.86872227580866]
  Few-shot image classification aims to learn an image classifier using only a small set of labeled examples per class. We propose DISEF, a novel approach which addresses the generalization challenge in few-shot learning using synthetic data. We validate our method in ten different benchmarks, consistently outperforming baselines and establishing a new state-of-the-art for few-shot classification.
  arXiv  Detail & Related papers  (2023-12-05T17:18:09Z)
• Contrastive Transformer Learning with Proximity Data Generation for Text-Based Person Search [60.626459715780605]
  Given a descriptive text query, text-based person search aims to retrieve the best-matched target person from an image gallery. Such a cross-modal retrieval task is quite challenging due to significant modality gap, fine-grained differences and insufficiency of annotated data. In this paper, we propose a simple yet effective dual Transformer model for text-based person search.
  arXiv  Detail & Related papers  (2023-11-15T16:26:49Z)
• Prompt Tuning Pushes Farther, Contrastive Learning Pulls Closer: A Two-Stage Approach to Mitigate Social Biases [13.837927115198308]
  We propose an adversarial training-inspired two-stage debiasing model using Contrastive learning and Continuous Prompt Augmentation. Our approach guides the model to achieve stronger debiasing performance by adding difficulty to the training process.
  arXiv  Detail & Related papers  (2023-07-04T09:35:03Z)
• Partner-Assisted Learning for Few-Shot Image Classification [54.66864961784989]
  Few-shot Learning has been studied to mimic human visual capabilities and learn effective models without the need of exhaustive human annotation. In this paper, we focus on the design of training strategy to obtain an elemental representation such that the prototype of each novel class can be estimated from a few labeled samples. We propose a two-stage training scheme, which first trains a partner encoder to model pair-wise similarities and extract features serving as soft-anchors, and then trains a main encoder by aligning its outputs with soft-anchors while attempting to maximize classification performance.
  arXiv  Detail & Related papers  (2021-09-15T22:46:19Z)
• Learning Fast Sample Re-weighting Without Reward Data [41.92662851886547]
  This paper presents a novel learning-based fast sample re-weighting (FSR) method that does not require additional reward data. Our experiments show the proposed method achieves competitive results compared to state of the arts on label noise and long-tailed recognition.
  arXiv  Detail & Related papers  (2021-09-07T17:30:56Z)
• Towards Model-Agnostic Post-Hoc Adjustment for Balancing Ranking Fairness and Algorithm Utility [54.179859639868646]
  Bipartite ranking aims to learn a scoring function that ranks positive individuals higher than negative ones from labeled data. There have been rising concerns on whether the learned scoring function can cause systematic disparity across different protected groups. We propose a model post-processing framework for balancing them in the bipartite ranking scenario.
  arXiv  Detail & Related papers  (2020-06-15T10:08:39Z)

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.