This study finds that the comprehensive development degree (CDD) of the finance subsystem is less fluctuated than that of the air environment subsystem, and both subsystems share similarities in spatial distributions. The coupling coordination degrees (CCD) keep fluctuating with varied development directions and extents in different regions; besides, the eastern regions are higher than the western ones for the coupling coordination degrees. In the next years, the coordination degrees of the regions will have different tendencies: despite of the former fluctuation trends, regions in the coordination range will have upward trends, while those in the transition range will be likely to decline. The results are useful in proposing corresponding measures to promote the coordination development between finance and air environment.

No abstract

No abstract

Training deep reinforcement learning (DRL) models usually requires high computation costs. Therefore, compressing DRL models possesses immense potential for training acceleration and model deployment. However, existing methods that generate small models mainly adopt the knowledge distillation based approach by iteratively training a dense network, such that the training process still demands massive computing resources. Indeed, sparse training from scratch in DRL has not been well explored and is particularly challenging due to non-stationarity in bootstrap training. In this work, we propose a novel sparse DRL training framework, "the Rigged Reinforcement Learning Lottery" (RLx2), which is capable of training a DRL agent using an ultra-sparse network throughout for off-policy reinforcement learning. The systematic RLx2 framework contains three key components: gradient-based topology evolution, multi-step Temporal Difference (TD) targets, and dynamic-capacity replay buffer. RLx2 enables efficient topology exploration and robust Q-value estimation simultaneously. We demonstrate state-of-the-art sparse training performance in several continuous control tasks using RLx2, showing 7.5×-20× model compression with less than 3% performance degradation, and up to 20× and 50× FLOPs reduction for training and inference, respectively.