Learning Population-Level Diffusions with Generative RNNs

Tatsunori Hashimoto, David Gifford, Tommi Jaakkola
Proceedings of The 33rd International Conference on Machine Learning, PMLR 48:2417-2426, 2016.

Abstract

We estimate stochastic processes that govern the dynamics of evolving populations such as cell differentiation. The problem is challenging since longitudinal trajectory measurements of individuals in a population are rarely available due to experimental cost and/or privacy. We show that cross-sectional samples from an evolving population suffice for recovery within a class of processes even if samples are available only at a few distinct time points. We provide a stratified analysis of recoverability conditions, and establish that reversibility is sufficient for recoverability. For estimation, we derive a natural loss and regularization, and parameterize the processes as diffusive recurrent neural networks. We demonstrate the approach in the context of uncovering complex cellular dynamics known as the ‘epigenetic landscape’ from existing biological assays.

Cite this Paper

BibTeX
@InProceedings{pmlr-v48-hashimoto16, title = {Learning Population-Level Diffusions with Generative RNNs}, author = {Hashimoto, Tatsunori and Gifford, David and Jaakkola, Tommi}, booktitle = {Proceedings of The 33rd International Conference on Machine Learning}, pages = {2417--2426}, year = {2016}, editor = {Balcan, Maria Florina and Weinberger, Kilian Q.}, volume = {48}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {20--22 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v48/hashimoto16.pdf}, url = {https://proceedings.mlr.press/v48/hashimoto16.html}, abstract = {We estimate stochastic processes that govern the dynamics of evolving populations such as cell differentiation. The problem is challenging since longitudinal trajectory measurements of individuals in a population are rarely available due to experimental cost and/or privacy. We show that cross-sectional samples from an evolving population suffice for recovery within a class of processes even if samples are available only at a few distinct time points. We provide a stratified analysis of recoverability conditions, and establish that reversibility is sufficient for recoverability. For estimation, we derive a natural loss and regularization, and parameterize the processes as diffusive recurrent neural networks. We demonstrate the approach in the context of uncovering complex cellular dynamics known as the ‘epigenetic landscape’ from existing biological assays.} }
Endnote
%0 Conference Paper %T Learning Population-Level Diffusions with Generative RNNs %A Tatsunori Hashimoto %A David Gifford %A Tommi Jaakkola %B Proceedings of The 33rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2016 %E Maria Florina Balcan %E Kilian Q. Weinberger %F pmlr-v48-hashimoto16 %I PMLR %P 2417--2426 %U https://proceedings.mlr.press/v48/hashimoto16.html %V 48 %X We estimate stochastic processes that govern the dynamics of evolving populations such as cell differentiation. The problem is challenging since longitudinal trajectory measurements of individuals in a population are rarely available due to experimental cost and/or privacy. We show that cross-sectional samples from an evolving population suffice for recovery within a class of processes even if samples are available only at a few distinct time points. We provide a stratified analysis of recoverability conditions, and establish that reversibility is sufficient for recoverability. For estimation, we derive a natural loss and regularization, and parameterize the processes as diffusive recurrent neural networks. We demonstrate the approach in the context of uncovering complex cellular dynamics known as the ‘epigenetic landscape’ from existing biological assays.
RIS
TY - CPAPER TI - Learning Population-Level Diffusions with Generative RNNs AU - Tatsunori Hashimoto AU - David Gifford AU - Tommi Jaakkola BT - Proceedings of The 33rd International Conference on Machine Learning DA - 2016/06/11 ED - Maria Florina Balcan ED - Kilian Q. Weinberger ID - pmlr-v48-hashimoto16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 48 SP - 2417 EP - 2426 L1 - http://proceedings.mlr.press/v48/hashimoto16.pdf UR - https://proceedings.mlr.press/v48/hashimoto16.html AB - We estimate stochastic processes that govern the dynamics of evolving populations such as cell differentiation. The problem is challenging since longitudinal trajectory measurements of individuals in a population are rarely available due to experimental cost and/or privacy. We show that cross-sectional samples from an evolving population suffice for recovery within a class of processes even if samples are available only at a few distinct time points. We provide a stratified analysis of recoverability conditions, and establish that reversibility is sufficient for recoverability. For estimation, we derive a natural loss and regularization, and parameterize the processes as diffusive recurrent neural networks. We demonstrate the approach in the context of uncovering complex cellular dynamics known as the ‘epigenetic landscape’ from existing biological assays. ER -
APA
Hashimoto, T., Gifford, D. & Jaakkola, T.. (2016). Learning Population-Level Diffusions with Generative RNNs. Proceedings of The 33rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 48:2417-2426 Available from https://proceedings.mlr.press/v48/hashimoto16.html.

Related Material