5 April 2023 in IEEE Transactions on Geoscience and Remote Sensing
Precipitation nowcasting, the high-resolution forecasting of precipitation in a short term, is essential in various applications in the real world. Previous deep learning methods use huge samples to learn potential laws, and the learning process lacks regularity, making it difficult to model the complex nonlinear precipitation phenomenon. Inspired by traditional numerical weather prediction models, we propose the MultiModal RNN (MM-RNN), which introduces knowledge of elements to guide precipitation prediction. This constraint forces the movement of precipitation to follow the underlying atmospheric motion laws. MM-RNN not only can provide accurate precipitation nowcasting but other meteorological elements predictions. Besides, it has high flexibility and is compatible with multiple RNN models, such as ConvLSTM, PredRNN, MIM, MotionRNN, etc. We conduct experiments on two multimodal datasets (MeteoNet and RAIN-F) and the results indicate that MM-RNN is superior to common RNN (MultiScale RNN, MS-RNN) using a single radar modality. For the MeteoNet, compared to MS-MotionRNN, the CSI (R ⩾ 10) of MM-MotionRNN increases by 23.4%, and the MSE of MM-MotionRNN decreases by 6.7%. For the RAIN-F, compared to MS-MIM, the HSS (R ⩾ 5) of MM-MIM increases by 209.4%, and the B-MSE of MM-MIM decreases by 4.6%.
13 February 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Precipitation nowcasting based on artificial intelligence has garnered widespread attention in the meteorological and computer communities in recent years. While new models are continuously proposed to refresh the forecasting performance, the problem of gradual blurring of forecast maps as the forecast period extends is still serious. Most models use the mean loss and the recursive prediction structure (such as MS-RNN). The mean loss always results in an average of future states, visually appearing as a blur. The recursive prediction method brings the accumulation of error (blur), causing the error (blur) of long-term predictions to increase exponentially. In this study, we add the adversarial loss and gradient loss to penalize the network to ease the blur caused by the averaging loss, and we introduce an additional deblurring network (composed of MS-RNN) behind the forecasting network (composed of MS-RNN) to alleviate the blur caused by the recursive structure, which reduces the blur of the current frame and then recursively and incrementally reduces the blur of subsequent frames. We name the proposed model DB-RNN, which can slow down the error accumulation and alleviate the blurring dilemma. Like MS-RNN, DB-RNN is compatible with multiple RNN models, such as ConvLSTM, TrajGRU, PredRNN, PredRNN++, MIM, MotionRNN, PrecipLSTM, etc. Experiments on two large radar datasets named HKO-7 and DWD-12 indicate that DB-RNN's predictions are more accurate and clear than those from MS-RNN.
11 August 2022 in IEEE Transactions on Geoscience and Remote Sensing
Accurate and timely nowcasting precipitation has huge social and economic benefits. However, the changes in clouds including expansion, dissipation, and distortion are extremely complex, which exacerbates the difficulty of forecasting. Fortunately, it still follows certain meteorological laws, which can be explored based on spatiotemporal information but are not fully considered by previous models. In this article, we design two modules to focus on these messages based on atmospheric characteristics. Specifically, the spatial local attention memory (SLAM) module combines local attention and memory mechanism to capture the meteorological spatial relationship, while the time difference memory (TDM) module combines differential technology and memory mechanism to capture the meteorological temporal variants. We combine these two modules with PredRNN and propose PrecipLSTM to sufficiently capture the spatiotemporal dependencies of radar data. We do exhaustive experiments with five baselines on four radar datasets. It is verified that PrecipLSTM achieves state-of-the-art results with fewer parameters than the previous state-of-the-art method.
1 March 2024 in IEEE Internet of Things Journal
Nowadays, mobile edge computing (MEC) appears as a new computing paradigm with its ability to utilize the computing power of both local devices and edge servers. In MEC, edge pricing and resource allocation (EPRA) are two important problems. Edge servers make a profit by selling computing services to users. To maximize their revenue, they need to determine an appropriate price for each user, and decide the amount of resources allocated to each user. However, none of the existing works consider the effect of users’ task assignment strategy on the revenue of the edge. In fact, EPRA will affect the users’ task offloading decision, as they expect to minimize their total cost. In turn, the users’ decision will also influence the revenue of the edge. Therefore, the interaction between mobile users and edge servers should be considered carefully and the interests of both sides need to be maximized simultaneously. In this article, we model the interaction between the two sides as a Stackelberg game. First, given a specified EPRA strategy, we derive a near-optimal task assignment strategy for each user to minimize the total cost based on a greedy algorithm UTA-G. Then, by applying the backward induction method, two pricing and resource allocation schemes with different granularity, i.e., user-granular EPRA and task-granular EPRA are proposed to bring higher revenue to the edge. Experimental results demonstrate that all the proposed algorithms can have good performance in task-intensive, resource-deficient, and workload-heavy scenarios.
10 August 2023 in IEEE Internet of Things Journal
With the rapid growth of the number of devices generating and collecting data, dispersion becomes an important feature of data in Internet of Things. Federated learning (FL) provides a feasible way to mine information in such distributed data. It involves training machine learning models over multiple distributed participants without raw data transmission. However, due to the data heterogeneity among participants, the performance of the FL model degrades dramatically. Currently, improved methods mainly reduce data heterogeneity from the perspective of modifying the process of model training, which usually have problems such as high resource consumption or the need for auxiliary data. In this paper, we enhance FL model from another perspective, focusing on data rather than model training. We reduce data heterogeneity by enhancing the trained local data to improve FL performance. Specifically, we propose an FL method based on data augmentation (abbr. FedM-UNE), implementing the classic data augmentation method MixUp in federated scenarios without transferring raw data. Furthermore, in order to adapt this method to regression tasks, we first modify MixUp by bilateral neighborhood expansion (MixUp-BNE), and then propose a federated data augmentation method named FedM-BNE based on it. Compared with the conventional FL method, both FedM-UNE and FedM-BNE increase negligible overhead. To demonstrate the effectiveness, we conduct exhaustive experiments on six datasets employing a variety of loss functions. The results indicate that FedM-UNE and FedM-BNE consistently improve the performance of the FL model. Moreover, our methods are compatible with existing FL enhancements, which yield further improvements in performance.
28 July 2022 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Precipitation nowcasting is an important but hard problem. Currently, with the landing of deep learning, it has been treated as an image prediction problem based on radar echo maps. However, deep learning models suffer from poor performance and blurred prediction results. Lots of improvement works enhance the model by adding complex modules, which increases insufferable training memory and time overhead. Others tempt to add more limitations or guidances on loss, but they usually have little effect in such an extremely complex and difficult task. In this article, we propose a simple but effective loss named focal frame loss (FFL), which assigns different weights to the images in the prediction sequence to focus on the images that are relatively difficult to predict. Experiments on two large-scale radar datasets show that FFL can greatly improve the performance of multiple popular models without introducing additional training costs.