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 2402.07087cs.LGcs.AIcs.CVstat.ML
收藏

生成模型训练的自校正自消耗循环

Self-Correcting Self-Consuming Loops for Generative Model Training

Brown University(布朗大学);Google DeepMind(谷歌DeepMind)

Nate Gillman, Michael Freeman, Daksh Aggarwal, Chia-Hong Hsu, Calvin Luo, Yonglong Tian, Chen Sun

AI总结:

本文针对生成模型训练中合成数据引发的自消耗循环不稳定问题,提出基于专家知识的自校正函数以逼近理想校正器,并在人体运动合成任务上验证其可在合成数据占比高达100%时避免模型崩溃。

摘要:

随着合成数据质量不断提高并在互联网上大量扩散,机器学习模型越来越多地在人类生成数据和机器生成数据的混合数据上进行训练。尽管利用合成数据进行表示学习已有成功案例,但将合成数据用于生成模型训练会产生“自消耗循环”,除非满足特定条件,否则可能导致训练不稳定甚至崩溃。本文旨在稳定自消耗生成模型训练。我们的理论结果表明,通过引入一个理想化的校正函数——该函数将数据点映射为在真实数据分布下更可能出现的形式——可以使自消耗循环的稳定性呈指数级提升。随后,我们提出了自校正函数,这些函数依赖专家知识(例如模拟器中编程实现的物理定律),旨在自动且大规模地逼近理想化校正器。我们在具有挑战性的人体运动合成任务上实证验证了自校正自消耗循环的有效性,并观察到即使合成数据与真实数据的比例高达100%,该方法也能成功避免模型崩溃。

英文摘要:

As synthetic data becomes higher quality and proliferates on the internet, machine learning models are increasingly trained on a mix of human- and machine-generated data. Despite the successful stories of using synthetic data for representation learning, using synthetic data for generative model training creates "self-consuming loops" which may lead to training instability or even collapse, unless certain conditions are met. Our paper aims to stabilize self-consuming generative model training. Our theoretical results demonstrate that by introducing an idealized correction function, which maps a data point to be more likely under the true data distribution, self-consuming loops can be made exponentially more stable. We then propose self-correction functions, which rely on expert knowledge (e.g. the laws of physics programmed in a simulator), and aim to approximate the idealized corrector automatically and at scale. We empirically validate the effectiveness of self-correcting self-consuming loops on the challenging human motion synthesis task, and observe that it successfully avoids model collapse, even when the ratio of synthetic data to real data is as high as 100%.

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。

↑