Future machine learning (ML) powered applications, such as autonomous driving and augmented reality, involve training and inference tasks with timeliness requirements and are communication and computation intensive, which demands for the edge learnin
g framework. The real-time requirements drive us to go beyond accuracy for ML. In this article, we introduce the concept of timely edge learning, aiming to achieve accurate training and inference while minimizing the communication and computation delay. We discuss key challenges and propose corresponding solutions from data, model and resource management perspectives to meet the timeliness requirements. Particularly, for edge training, we argue that the total training delay rather than rounds should be considered, and propose data or model compression, and joint device scheduling and resource management schemes for both centralized training and federated learning systems. For edge inference, we explore the dependency between accuracy and delay for communication and computation, and propose dynamic data compression and flexible pruning schemes. Two case studies show that the timeliness performances, including the training accuracy under a given delay budget and the completion ratio of inference tasks within deadline, are highly improved with the proposed solutions.
In federated learning (FL), devices contribute to the global training by uploading their local model updates via wireless channels. Due to limited computation and communication resources, device scheduling is crucial to the convergence rate of FL. In
this paper, we propose a joint device scheduling and resource allocation policy to maximize the model accuracy within a given total training time budget for latency constrained wireless FL. A lower bound on the reciprocal of the training performance loss, in terms of the number of training rounds and the number of scheduled devices per round, is derived. Based on the bound, the accuracy maximization problem is solved by decoupling it into two sub-problems. First, given the scheduled devices, the optimal bandwidth allocation suggests allocating more bandwidth to the devices with worse channel conditions or weaker computation capabilities. Then, a greedy device scheduling algorithm is introduced, which in each step selects the device consuming the least updating time obtained by the optimal bandwidth allocation, until the lower bound begins to increase, meaning that scheduling more devices will degrade the model accuracy. Experiments show that the proposed policy outperforms state-of-the-art scheduling policies under extensive settings of data distributions and cell radius.
Owing to the increasing need for massive data analysis and model training at the network edge, as well as the rising concerns about the data privacy, a new distributed training framework called federated learning (FL) has emerged. In each iteration o
f FL (called round), the edge devices update local models based on their own data and contribute to the global training by uploading the model updates via wireless channels. Due to the limited spectrum resources, only a portion of the devices can be scheduled in each round. While most of the existing work on scheduling focuses on the convergence of FL w.r.t. rounds, the convergence performance under a total training time budget is not yet explored. In this paper, a joint bandwidth allocation and scheduling problem is formulated to capture the long-term convergence performance of FL, and is solved by being decoupled into two sub-problems. For the bandwidth allocation sub-problem, the derived optimal solution suggests to allocate more bandwidth to the devices with worse channel conditions or weaker computation capabilities. For the device scheduling sub-problem, by revealing the trade-off between the number of rounds required to attain a certain model accuracy and the latency per round, a greedy policy is inspired, that continuously selects the device that consumes the least time in model updating until achieving a good trade-off between the learning efficiency and latency per round. The experiments show that the proposed policy outperforms other state-of-the-art scheduling policies, with the best achievable model accuracy under training time budgets.
