arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1804.01523cs.CVcs.AIcs.LGcs.RO
收藏

随机对抗视频预测

Stochastic Adversarial Video Prediction

University of California, Berkeley(加利福尼亚大学伯克利分校)

Alex X. Lee, Richard Zhang, Frederik Ebert, Pieter Abbeel, Chelsea Finn, Sergey Levine

摘要:

要想预测未来可能发生的事情,就需要深入了解支配世界的物理和因果规则。能够做到这一点的模型有许多吸引人的应用,从机器人规划到表示学习。然而,学习预测原始的未来观察结果(例如视频中的帧)是非常具有挑战性的-问题的模糊性可能导致一个天真设计的模型将可能的未来平均为一个单一的模糊预测。最近,这已经通过两种不同的方法来解决:(a)显式建模潜在随机性的潜在变分变量模型和(b)旨在产生自然图像的逆向训练模型。然而,标准的潜变量模型很难产生现实的结果,而标准的逆向训练模型没有充分利用潜变量,无法产生多样化的预测。我们表明,这些不同的方法实际上是互补的。将两者结合起来,可以产生对人类评级者来说更现实的预测,并更好地涵盖可能的未来范围。我们的方法在这些方面优于先前和并发的工作。

英文摘要:

Being able to predict what may happen in the future requires an in-depth understanding of the physical and causal rules that govern the world. A model that is able to do so has a number of appealing applications, from robotic planning to representation learning. However, learning to predict raw future observations, such as frames in a video, is exceedingly challenging -- the ambiguous nature of the problem can cause a naively designed model to average together possible futures into a single, blurry prediction. Recently, this has been addressed by two distinct approaches: (a) latent variational variable models that explicitly model underlying stochasticity and (b) adversarially-trained models that aim to produce naturalistic images. However, a standard latent variable model can struggle to produce realistic results, and a standard adversarially-trained model underutilizes latent variables and fails to produce diverse predictions. We show that these distinct methods are in fact complementary. Combining the two produces predictions that look more realistic to human raters and better cover the range of possible futures. Our method outperforms prior and concurrent work in these aspects.

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。

↑