Large‐scale knowledge distillation with elastic heterogeneous computing resources

Abstract

Although more layers and more parameters generally improve the accuracy of the models, such big models generally have high computational complexity and require big memory, which exceed the capacity of small devices for inference and incurs long training time. In addition, it is difficult to afford long training time and inference time of big models even in high performance servers, as well. As an efficient approach to compress a large deep model (a teacher model) to a compact model (a student model), knowledge distillation emerges as a promising approach to deal with the big models. Existing knowledge distillation methods cannot exploit the elastic available computing resources and correspond to low efficiency. In this paper, we propose an Elastic Deep Learning framework for knowledge Distillation, that is, EDL‐Dist. The advantages of EDL‐Dist are threefold. First, the inference and the training process is separated. Second, elastic available computing resources can be utilized to improve the efficiency. Third, fault‐tolerance of the training and inference processes is supported. We take extensive experimentation to show that the throughput of EDL‐Dist is up to 3.125 times faster than the baseline method (online knowledge distillation) while the accuracy is similar or higher.

Multi-Job Intelligent Scheduling With Cross-Device Federated Learning

“…FL only allows the intermediate data to be transferred from the distributed devices, which can be the weights or the gradients of a model. FL generally utilizes a parameter server architecture [9], [10], [11], where a server (or a group of servers) coordinates the training process with numerous devices. To collaboratively train a global model, the server selects (schedules) several devices to perform local model updates based on their local data, and then it aggregates the local models to obtain a new global model.…”

Distributed and deep vertical federated learning with big data

“…We apply the parameter server and peer‐to‐peer communication architecture to accelerate the big data processing. The proposed architecture can be deployed in a large‐scale cluster environment environment or a cloud environment, 26 where multiple powerful servers can be offered for parallel computing 27 …”

“…The proposed architecture can be deployed in a large-scale cluster environment environment or a cloud environment, 26 where multiple powerful servers can be offered for parallel computing. 27 We summarize our main contributions of this article as follows:…”