L1 regularization is better than L2 for learning and predicting chaotic systems
Z. Szabo, A. Lorincz
Abstract
Emergent behaviors are in the focus of recent research interest. It is then of considerable importance to investigate what optimizations suit the learning and prediction of chaotic systems, the putative candidates for emergence. We have compared L1 and L2 regularizations on predicting chaotic time series using linear recurrent neural networks. The internal representation and the weights of the networks were optimized in a unifying framework. Computational tests on different problems indicate considerable advantages for the L1 regularization: It had considerably better learning time and better interpolating capabilities. We shall argue that optimization viewed as a maximum likelihood estimation justifies our results, because L1 regularization fits heavy-tailed distributions -- an apparently general feature of emergent systems -- better.
Create a lesson
Related papers
Coarse Semantic Blocks for Efficient Training-Free Sparse Attention
Sosuke Kobayashi, Shoichiro Yamaguchi, Seiya Tokui
What Is the Influence of Task Order in Continual Learning? An Empirical Study and Order-Aware Evaluation for Motion Forecasting
Zirui Li, Yunlong Lin, Haohan Yang et al.
Reciprocity Can Erase Direction:tructural Boundaries and Controlled Utility of Magnetic Graph Networks
Dawei Han, Zicheng Wang, Yaoxing Ren et al.
Deep Learning Compilers: Architecture, Optimization, and Empirical Characterization
Jinghao Wen, Ruixuan Wang, Xun Jiao
OmniST: Topology-Invariant Spatiotemporal Forecasting with Guaranteed-Coverage Uncertainty at Unseen Sensor Locations
Praveg Vashishtha
LIRA: A More Interpretable Dynamic Loss-Informed Rank Allocation for Low-Rank Adaptation
Sihao Wang, Bozheng Li, Xuemiao Xu et al.