ABSTRACT
The capacity to anticipate streamflow is critical to the efficient functioning of reservoir systems as it gives vital information to reservoir operators about water release quantities as well as help quantify the impact of environmental factors on downstream water quality. Yet, streamflow modelling is difficult owing to the intricate interactions between different watershed outlets. In this paper, we argue that one possible solution to this problem is to identify the causal structure of these outlets, which would allow for the identification of crucial watershed outlets while capturing the spatiotemporally informed complex relationships leading to improved hydrological resource management. However, due to the inherent complexity of spatiotemporal causal learning problems, extending existing causal discovery methods to a whole basin is a major hurdle. To address these issues, we offer STREAMS, a new framework that uses Reinforcement Learning (RL) to optimize the search space for causal discovery and an LSTM-GCN based autoencoder to infer spatiotemporal causal features for streamflow rate prediction. We conduct extensive experiments on the Brazos river basin carried out within the scope of a US Army Corps of Engineers, Engineering With Nature Initiative project, including empirical studies of generalization performance to verify the nature of the inferred relationships.
Supplemental Material
- H Assaf and M Saadeh. 2009. Geostatistical assessment of groundwater nitrate contamination with reflection on DRASTIC vulnerability assessment: the case of the Upper Litani Basin, Lebanon. Water resources management, Vol. 23 (2009), 775--796.Google ScholarCross Ref
- Peter Bühlmann. 2020. Invariance, causality and robustness. Statist. Sci. (2020).Google Scholar
- A Casado, S Bermudo, AD López-Sánchez, and J Sánchez-Oro. 2023. An iterated greedy algorithm for finding the minimum dominating set in graphs. Mathematics and Computers in Simulation, Vol. 207 (2023), 41--58.Google ScholarDigital Library
- Tianfeng Chai and Roland R Draxler. 2014. Root mean square error (RMSE) or mean absolute error (MAE)?--Arguments against avoiding RMSE in the literature. Geoscientific model development, Vol. 7, 3 (2014), 1247--1250.Google Scholar
- Hyeon Seok Choi, Joong Hoon Kim, Eui Hoon Lee, and Sun-Kwon Yoon. 2022. Development of a Revised Multi-Layer Perceptron Model for Dam Inflow Prediction. Water, Vol. 14, 12 (2022), 1878.Google ScholarCross Ref
- Hamidreza Ghasemi Damavandi, Reepal Shah, Dimitrios Stampoulis, Yuhang Wei, Dragan Boscovic, John Sabo, et al. 2019. Accurate prediction of streamflow using long short-term memory network: a case study in the Brazos River Basin in Texas. International Journal of Environmental Science and Development, Vol. 10, 10 (2019), 294--300.Google ScholarCross Ref
- Arnaud De Myttenaere, Boris Golden, Bénédicte Le Grand, and Fabrice Rossi. 2016. Mean absolute percentage error for regression models. Neurocomputing, Vol. 192 (2016), 38--48.Google ScholarCross Ref
- Philip Hans Franses. 2016. A note on the mean absolute scaled error. International Journal of Forecasting, Vol. 32, 1 (2016), 20--22.Google ScholarCross Ref
- Dominique MA Haughton. 1988. On the choice of a model to fit data from an exponential family. The annals of statistics (1988), 342--355.Google Scholar
- HAWQS. 2020. HAWQS System and Data to Model the Lower 48 Conterminous U.S using the SWAT Model. Dataset. https://doi.org/10.18738/T8/XN3TE0Google Scholar
- Yiyi Huang, Matth"aus Kleindessner, Alexey Munishkin, Debvrat Varshney, Pei Guo, and Jianwu Wang. 2021. Benchmarking of data-driven causality discovery approaches in the interactions of arctic sea ice and atmosphere. Frontiers in big Data, Vol. 4 (2021), 642182.Google Scholar
- Dostdar Hussain, Tahir Hussain, Aftab Ahmed Khan, Syed Ali Asad Naqvi, and Akhtar Jamil. 2020. A deep learning approach for hydrological time-series prediction: A case study of Gilgit river basin. Earth Science Informatics, Vol. 13 (2020), 915--927.Google ScholarCross Ref
- Piraporn Jangyodsuk, Dong-Jun Seo, and Jean Gao. 2014. Causal graph discovery for hydrological time series knowledge discovery. International Conference on Hydroinformatics (2014).Google Scholar
- Leslie Pack Kaelbling, Michael L Littman, and Andrew W Moore. 1996. Reinforcement learning: A survey. Journal of artificial intelligence research, Vol. 4 (1996), 237--285.Google ScholarDigital Library
- Jaewon Kwak, Heechan Han, Soojun Kim, and Hung Soo Kim. 2022. Is the deep-learning technique a completely alternative for the hydrological model?: A case study on Hyeongsan River Basin, Korea. Stochastic Environmental Research and Risk Assessment (2022), 1--15.Google ScholarCross Ref
- Eui Hoon Lee. 2022. Runoff prediction of urban stream based on the discharge of pump stations using improved multi-layer perceptron applying new optimizers combined with a harmony search. Journal of Hydrology, Vol. 615 (2022), 128708.Google ScholarCross Ref
- Alexey Natekin and Alois Knoll. 2013. Gradient boosting machines, a tutorial. Frontiers in neurorobotics, Vol. 7 (2013), 21.Google Scholar
- Meike Nauta, Doina Bucur, and Christin Seifert. 2019. Causal discovery with attention-based convolutional neural networks. Machine Learning and Knowledge Extraction, Vol. 1, 1 (2019), 312--340.Google ScholarCross Ref
- Behnam Neyshabur, Srinadh Bhojanapalli, David McAllester, and Nati Srebro. 2017. Exploring generalization in deep learning. Advances in neural information processing systems, Vol. 30 (2017).Google Scholar
- Mohammed Ombadi, Phu Nguyen, Soroosh Sorooshian, and Kuo-lin Hsu. 2020. Evaluation of methods for causal discovery in hydrometeorological systems. Water Resources Research, Vol. 56, 7 (2020), e2020WR027251.Google ScholarCross Ref
- Roxana Pamfil, Nisara Sriwattanaworachai, Shaan Desai, Philip Pilgerstorfer, Konstantinos Georgatzis, Paul Beaumont, and Bryon Aragam. 2020. Dynotears: Structure learning from time-series data. In International Conference on Artificial Intelligence and Statistics. PMLR, 1595--1605.Google Scholar
- Mark R Segal. 2004. Machine learning benchmarks and random forest regression. (2004).Google Scholar
- Paras Sheth, Ting Liu, Durmus Doner, Qi Deng, Yuhang Wei, Rebecca Muenich, John Sabo, K Selcc uk Candan, and Huan Liu. 2022a. Causal Discovery for Feature Selection in Physical Process-Based Hydrological Systems. In 2022 IEEE International Conference on Big Data (Big Data). IEEE, 5568--5577.Google Scholar
- Paras Sheth, Raha Moraffah, K Selcc uk Candan, Adrienne Raglin, and Huan Liu. 2022b. Domain Generalization--A Causal Perspective. arXiv preprint arXiv:2209.15177 (2022).Google Scholar
- Paras Sheth, Reepal Shah, John Sabo, K Selcc uk Candan, and Huan Liu. 2022c. STCD: A Spatio-Temporal Causal Discovery Framework for Hydrological Systems. In 2022 IEEE International Conference on Big Data (Big Data). IEEE, 5578--5583.Google Scholar
- David Silver, Guy Lever, Nicolas Heess, Thomas Degris, Daan Wierstra, and Martin Riedmiller. 2014. Deterministic policy gradient algorithms. In International conference on machine learning. Pmlr, 387--395.Google Scholar
- Ralf C Staudemeyer and Eric Rothstein Morris. 2019. Understanding LSTM--a tutorial into long short-term memory recurrent neural networks. arXiv preprint arXiv:1909.09586 (2019).Google Scholar
- Xinwei Sun, Botong Wu, Xiangyu Zheng, Chang Liu, Wei Chen, Tao Qin, and Tie-Yan Liu. 2021. Recovering latent causal factor for generalization to distributional shifts. Advances in Neural Information Processing Systems, Vol. 34 (2021), 16846--16859.Google Scholar
- Manoj Tiwaskar, Yash Garg, Xinsheng Li, K Selcc uk Candan, and Maria Luisa Sapino. 2021. Selego: robust variate selection for accurate time series forecasting. Data Mining and Knowledge Discovery, Vol. 35 (2021), 2141--2167.Google ScholarDigital Library
- Chris Wallace, Kevin B Korb, and Honghua Dai. 1996. Causal discovery via MML. In ICML, Vol. 96. 516--524.Google Scholar
- Sifan Wang, Shyam Sankaran, and Paris Perdikaris. 2022. Respecting causality is all you need for training physics-informed neural networks. arXiv preprint arXiv:2203.07404 (2022).Google Scholar
- Chixuan Wei, Zhihai Wang, Jidong Yuan, Chuanming Li, and Shengbo Chen. 2023. Time-frequency based multi-task learning for semi-supervised time series classification. Information Sciences, Vol. 619 (2023), 762--780.Google ScholarDigital Library
- Santiago Zazo, José-Luis Molina, Verónica Ruiz-Ortiz, Mercedes Vélez-Nicolás, and Santiago Garcia-López. 2020. Modeling river runoff temporal behavior through a hybrid causal--hydrological (HCH) method. Water, Vol. 12, 11 (2020), 3137.Google ScholarCross Ref
- Shengyu Zhu, Ignavier Ng, and Zhitang Chen. 2019. Causal discovery with reinforcement learning. arXiv preprint arXiv:1906.04477 (2019).Google Scholar
Index Terms
- STREAMS: Towards Spatio-Temporal Causal Discovery with Reinforcement Learning for Streamflow Rate Prediction
Recommendations
Causal Discovery from Temporal Data
KDD '23: Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data MiningTemporal data representing chronological observations of complex systems can be ubiquitously collected in smart industry, medicine, finance and etc. In the last decade, many tasks have been studied for mining temporal data and offered significant value ...
Causal Graph Discovery for Explainable Insights on Marine Biotoxin Shellfish Contamination
Intelligent Data Engineering and Automated Learning – IDEAL 2023AbstractHarmful algal blooms are natural phenomena that cause shellfish contamination due to the rapid accumulation of marine biotoxins. To prevent public health risks, the Portuguese Institute of the Ocean and the Atmosphere (IPMA) regularly monitors ...
CORE: Towards Scalable and Efficient Causal Discovery with Reinforcement Learning
AAMAS '24: Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent SystemsCausal discovery is the challenging task of inferring causal structure from data. Motivated by Pearl's Causal Hierarchy (PCH), which tells us that passive observations alone are not enough to distinguish correlation from causation, there has been a ...
Comments