LEARNING GENERAL TRANSFORMATIONS OF DATA FOR OUT-OF-SAMPLE EXTENSIONS.

LEARNING GENERAL TRANSFORMATIONS OF DATA FOR OUT-OF-SAMPLE EXTENSIONS.
复制标题

DOI:
10.1109/mlsp49062.2020.9231660
复制
发表时间:
2020-09
期刊:
IEEE International Workshop on Machine Learning for Signal Processing : [proceedings]. IEEE International Workshop on Machine Learning for Signal Processing
影响因子:
--
通讯作者:
Krishnaswamy S
Krishnaswamy S
中科院分区:
其他
文献类型:
--
作者:
Amodio M;van Dijk D;Wolf G;Krishnaswamy S

文献摘要

参考文献

相似文献

虽然像Gans这样的生成性模型已经成功地从噪声映射到特定的数据分布,或者更一般地从一个数据分布映射到另一个数据分布,但它们不能隔离正在发生的变换并将其应用于训练中未见的新分布。因此,它们记住了变换的域,并且不能从样本中概括变换。为了解决这个问题,我们提出了一种新的神经网络,称为神经元变换网络(NTNet),它将代表变换本身的信号与代表内部分布变化的其他信号隔离开来。然后,可以从与原始训练数据分布不同的新数据集中移除该信号。我们在十几个合成和生物医学单细胞RNA测序数据集上展示了我们的NTNet的有效性,在这些数据集中,NTNet能够了解一个细胞样本上的遗传和药物扰动执行的数据转换,并成功地将其应用于另一个细胞样本以预测治疗结果。
While generative models such as GANs have been successful at mapping from noise to specific distributions of data, or more generally from one distribution of data to another, they cannot isolate the transformation that is occurring and apply it to a new distribution not seen in training. Thus, they memorize the domain of the transformation, and cannot generalize the transformation out of sample. To address this, we propose a new neural network called a Neuron Transformation Network (NTNet) that isolates the signal representing the transformation itself from the other signals representing internal distribution variation. This signal can then be removed from a new dataset distributed differently from the original one trained on. We demonstrate the effectiveness of our NTNet on more than a dozen synthetic and biomedical single-cell RNA sequencing datasets, where the NTNet is able to learn the data transformation performed by genetic and drug perturbations on one sample of cells and successfully apply it to another sample of cells to predict treatment outcome.
DOI: 10.1038/s41592-019-0576-7
发表时间: 2019-11
期刊: NATURE METHODS
影响因子: 48
作者:
Amodio, Matthew;van Dijk, David;Srinivasan, Krishnan;Chen, William S.;Mohsen, Hussein;Moon, Kevin R.;Campbell, Allison;Zhao, Yujiao;Wang, Xiaomei;Venkataswamy, Manjunatha;Desai, Anita;Ravi, V.;Kumar, Priti;Montgomery, Ruth;Wolf, Guy;Krishnaswamy, Smita
通讯作者: Krishnaswamy, Smita
DOI: 10.1038/s41592-019-0494-8
发表时间: 2019-08-01
期刊: NATURE METHODS
影响因子: 48
作者:
Lotfollahi, Mohammad;Wolf, F. Alexander;Theis, Fabian J.
通讯作者: Theis, Fabian J.
DOI: 10.1126/science.aan6828
发表时间: 2017-10-06
期刊: Science (New York, N.Y.)
影响因子: --
作者:
Stubbington MJT;Rozenblatt-Rosen O;Regev A;Teichmann SA
通讯作者: Teichmann SA
DOI: 10.1038/nature24489
发表时间: 2017-11-16
期刊: Nature
影响因子: 64.8
作者:
Haber AL;Biton M;Rogel N;Herbst RH;Shekhar K;Smillie C;Burgin G;Delorey TM;Howitt MR;Katz Y;Tirosh I;Beyaz S;Dionne D;Zhang M;Raychowdhury R;Garrett WS;Rozenblatt-Rosen O;Shi HN;Yilmaz O;Xavier RJ;Regev A
通讯作者: Regev A
使用自然遗传变异的多重液滴单细胞RNA序列。
DOI: 10.1038/nbt.4042
发表时间: 2018-01
影响因子: 46.9
作者:
Kang HM;Subramaniam M;Targ S;Nguyen M;Maliskova L;McCarthy E;Wan E;Wong S;Byrnes L;Lanata CM;Gate RE;Mostafavi S;Marson A;Zaitlen N;Criswell LA;Ye CJ
通讯作者: Ye CJ