Robust data sampling in machine learning: a game-theoretic framework for training and validation data selection. (English) Zbl 07865416
Summary: How to sample training/validation data is an important question for machine learning models, especially when the dataset is heterogeneous and skewed. In this paper, we propose a data sampling method that robustly selects training/validation data. We formulate the training/validation data sampling process as a two-player game: a trainer aims to sample training data so as to minimize the test error, while a validator adversarially samples validation data that can increase the test error. Robust sampling is achieved at the game equilibrium. To accelerate the searching process, we adopt reinforcement learning aided Monte Carlo trees search (MCTS). We apply our method to a car-following modeling problem, a complicated scenario with heterogeneous and random human driving behavior. Real-world data, the Next Generation SIMulation (NGSIM), is used to validate this method, and experiment results demonstrate the sampling robustness and thereby the model out-of-sample performance.
MSC:
| 68T05 | Learning and adaptive systems in artificial intelligence |
| 91A80 | Applications of game theory |
References:
| [1] | Fu, Y.; Xiang, T.; Jiang, Y.G.; Xue, X.; Sigal, L.; Gong, S.; Recent advances in zero-shot recognition: Toward data-efficient understanding of visual content; IEEE Signal Process. Mag.: 2018; Volume 35 ,112-125. |
| [2] | Mo, Z.; Shi, R.; Di, X.; A physics-informed deep learning paradigm for car-following models; Transp. Res. Part C Emerg. Technol.: 2021; Volume 130 ,103240. |
| [3] | Mo, Z.; Fu, Y.; TrafficFlowGAN: Physics-informed Flow based Generative Adversarial Network for Uncertainty Quantification; Proceedings of the European Conference on Machine Learning and Data Mining (ECML PKDD): ; . |
| [4] | Shi, R.; Mo, Z.; Huang, K.; Di, X.; Du, Q.; A physics-informed deep learning paradigm for traffic state and fundamental diagram estimation; IEEE Trans. Intell. Transp. Syst.: 2021; Volume 23 ,11688-11698. |
| [5] | Shi, R.; Mo, Z.; Di, X.; Physics-informed deep learning for traffic state estimation: A hybrid paradigm informed by second-order traffic models; Proceedings of the AAAI Conference on Artificial Intelligence: ; Volume Volume 35 ,540-547. |
| [6] | Ossen, S.; Hoogendoorn, S.P.; Validity of trajectory-based calibration approach of car-following models in presence of measurement errors; Transp. Res. Rec.: 2008; Volume 2088 ,117-125. |
| [7] | Hoogendoorn, S.; Hoogendoorn, R.; Calibration of microscopic traffic-flow models using multiple data sources; Philos. Trans. R. Soc. A Math. Phys. Eng. Sci.: 2010; Volume 368 ,4497-4517. · Zbl 1202.90065 |
| [8] | Fernández, A.; Garcia, S.; Herrera, F.; Chawla, N.V.; SMOTE for learning from imbalanced data: Progress and challenges, marking the 15-year anniversary; J. Artif. Intell. Res.: 2018; Volume 61 ,863-905. · Zbl 1443.68147 |
| [9] | Tokdar, S.T.; Kass, R.E.; Importance sampling: A review; Wiley Interdiscip. Rev. Comput. Stat.: 2010; Volume 2 ,54-60. |
| [10] | Kantarcıoğlu, M.; Xi, B.; Clifton, C.; Classifier evaluation and attribute selection against active adversaries; Data Min. Knowl. Discov.: 2011; Volume 22 ,291-335. · Zbl 1235.62067 |
| [11] | Liu, X.; Hsieh, C.J.; From adversarial training to generative adversarial networks; arXiv: 2018; . |
| [12] | Liu, G.; Khalil, I.; Khreishah, A.; GanDef: A GAN based Adversarial Training Defense for Neural Network Classifier; Proceedings of the IFIP International Conference on ICT Systems Security and Privacy Protection: Berlin/Heidelberg, Germany 2019; ,19-32. |
| [13] | Browne, C.B.; Powley, E.; Whitehouse, D.; Lucas, S.M.; Cowling, P.I.; Rohlfshagen, P.; Tavener, S.; Perez, D.; Samothrakis, S.; Colton, S.; A survey of monte carlo tree search methods; IEEE Trans. Comput. Intell. AI Games: 2012; Volume 4 ,1-43. |
| [14] | Zhou, M.; Qu, X.; Li, X.; A recurrent neural network based microscopic car following model to predict traffic oscillation; Transp. Res. Part C Emerg. Technol.: 2017; Volume 84 ,245-264. |
| [15] | Sharma, A.; Zheng, Z.; Bhaskar, A.; Is more always better? The impact of vehicular trajectory completeness on car-following model calibration and validation; Transp. Res. Part B Methodol.: 2019; Volume 120 ,49-75. |
| [16] | Wang, X.; Jiang, R.; Li, L.; Lin, Y.; Zheng, X.; Wang, F.Y.; Capturing car-following behaviors by deep learning; IEEE Trans. Intell. Transp. Syst.: 2017; Volume 19 ,910-920. |
| [17] | Zhu, M.; Wang, X.; Wang, Y.; Human-like autonomous car-following model with deep reinforcement learning; Transp. Res. Part C Emerg. Technol.: 2018; Volume 97 ,348-368. |
| [18] | Nageshrao, S.; Tseng, E.; Filev, D.; Autonomous Highway Driving using Deep Reinforcement Learning; arXiv: 2019; . |
| [19] | Silver, D.; Schrittwieser, J.; Simonyan, K.; Antonoglou, I.; Huang, A.; Guez, A.; Hubert, T.; Baker, L.; Lai, M.; Bolton, A.; Mastering the game of go without human knowledge; Nature: 2017; Volume 550 ,354. |
| [20] | Wang, K.; Sun, W.; Du, Q.; A cooperative game for automated learning of elasto-plasticity knowledge graphs and models with AI-guided experimentation; Comput. Mech.: 2019; Volume 64 ,467-499. · Zbl 1464.74034 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
