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 1906.01550stat.MLcs.LG
收藏

面向任务和结构独立的泛化间隙预测器

Towards Task and Architecture-Independent Generalization Gap Predictors

Google Research(谷歌研究院)

Scott Yak, Javier Gonzalvo, Hanna Mazzawi

摘要:

我们可以使用深度学习来预测深度学习何时起作用吗?我们的结果表明肯定的。我们通过训练13,500个具有不同架构的神经网络,在不同的螺旋数据集上,并使用不同的优化参数来创建数据集。我们使用该数据集为这些神经网络训练任务无关和架构无关的泛化间隙预测器。我们扩展了Jiang等人。(2018)也使用DNN和RNN,并证明它们优于线性模型,获得了$R^2=0.965$。我们还展示了独立于架构,独立于任务和分布泛化差距预测任务的结果。DNN和RNN都一致且显着优于线性模型,RNN获得$R^2=0.584$。

英文摘要:

Can we use deep learning to predict when deep learning works? Our results suggest the affirmative. We created a dataset by training 13,500 neural networks with different architectures, on different variations of spiral datasets, and using different optimization parameters. We used this dataset to train task-independent and architecture-independent generalization gap predictors for those neural networks. We extend Jiang et al. (2018) to also use DNNs and RNNs and show that they outperform the linear model, obtaining $R^2=0.965$. We also show results for architecture-independent, task-independent, and out-of-distribution generalization gap prediction tasks. Both DNNs and RNNs consistently and significantly outperform linear models, with RNNs obtaining $R^2=0.584$.

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。

↑