arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2302.12192cs.LGcs.AIcs.CV
收藏

使用人类反馈对齐文本到图像模型

Aligning Text-to-Image Models using Human Feedback

Google Research(谷歌研究院);University of California, Berkeley(加州大学伯克利分校)

Kimin Lee, Hao Liu, Moonkyung Ryu, Olivia Watkins, Yuqing Du, Craig Boutilier, Pieter Abbeel, Mohammad Ghavamzadeh, Shixiang Shane Gu

摘要:

深度生成模型在文本到图像的合成中显示出令人印象深刻的结果。然而,当前的文本到图像模型通常生成与文本提示不充分对齐的图像。我们提出了一种微调方法,用于调整这样的模型,使用人类的反馈,包括三个阶段。首先,我们从一组不同的文本提示中收集评估模型输出对齐的人类反馈。然后,我们使用人类标记的图像-文本数据集来训练预测人类反馈的奖励函数。最后,通过最大化奖励加权似然来微调文本到图像模型,以改善图文对齐。我们的方法比预先训练的模型更准确地生成具有指定颜色,计数和背景的对象。我们还分析了几种设计选择,并发现仔细调查,这样的设计选择是很重要的,在平衡的保真度的权衡。我们的研究结果证明了从人类反馈中学习的潜力,以显着改善文本到图像的模型。

英文摘要:

Deep generative models have shown impressive results in text-to-image synthesis. However, current text-to-image models often generate images that are inadequately aligned with text prompts. We propose a fine-tuning method for aligning such models using human feedback, comprising three stages. First, we collect human feedback assessing model output alignment from a set of diverse text prompts. We then use the human-labeled image-text dataset to train a reward function that predicts human feedback. Lastly, the text-to-image model is fine-tuned by maximizing reward-weighted likelihood to improve image-text alignment. Our method generates objects with specified colors, counts and backgrounds more accurately than the pre-trained model. We also analyze several design choices and find that careful investigations on such design choices are important in balancing the alignment-fidelity tradeoffs. Our results demonstrate the potential for learning from human feedback to significantly improve text-to-image models.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑