DNN Partitioning for Inference Throughput Acceleration at the Edge

Deep neural network (DNN) inference on streaming data requires computing resources to satisfy inference throughput requirements. However, latency and privacy sensitive deep learning applications cannot afford to offload computation to remote clouds because of the implied transmission cost and lack of trust in third-party cloud providers. Among solutions to increase performance while keeping computation on a constrained environment, hardware acceleration can be onerous, and model optimization requires extensive design efforts while hindering accuracy. DNN partitioning is a third complementary approach, and consists of distributing the inference workload over several available edge devices, taking into account the edge network properties and the DNN structure, with the objective of maximizing the inference throughput (number of inferences per second). This paper introduces a method to predict inference and transmission latencies for multi-threaded distributed DNN deployments, and defines an optimization process to maximize the inference throughput. A branch and bound solver is then presented and analyzed to quantify the achieved performance and complexity. This analysis has led to the definition of the acceleration region, which describes deterministic conditions on the DNN and network properties under which DNN partitioning is beneficial. Finally, experimental results confirm the simulations and show inference throughput improvements in sample edge deployments.

View this article on IEEE Xplore

 

Security Hardening of Intelligent Reflecting Surfaces Against Adversarial Machine Learning Attacks

Next-generation communication networks, also known as NextG or 5G and beyond, are the future data transmission systems that aim to connect a large amount of Internet of Things (IoT) devices, systems, applications, and consumers at high-speed data transmission and low latency. Fortunately, NextG networks can achieve these goals with advanced telecommunication, computing, and Artificial Intelligence (AI) technologies in the last decades and support a wide range of new applications. Among advanced technologies, AI has a significant and unique contribution to achieving these goals for beamforming, channel estimation, and Intelligent Reflecting Surfaces (IRS) applications of 5G and beyond networks. However, the security threats and mitigation for AI-powered applications in NextG networks have not been investigated deeply in academia and industry due to being new and more complicated. This paper focuses on an AI-powered IRS implementation in NextG networks along with its vulnerability against adversarial machine learning attacks. This paper also proposes the defensive distillation mitigation method to defend and improve the robustness of the AI-powered IRS model, i.e., reduce the vulnerability. The results indicate that the defensive distillation mitigation method can significantly improve the robustness of AI-powered models and their performance under an adversarial attack.

View this article on IEEE Xplore

 

Tool Wear Monitoring Based on Transfer Learning and Improved Deep Residual Network

Considering the complex structure weight of the existing tool wear state monitoring model based on deep learning, prone to over-fitting and requiring a large amount of training data, a monitoring method based on Transfer Learning and Improved Deep Residual Network is proposed. First, the data is preprocessed, one-dimensional cutting force data are transformed into two-dimensional spectrum by wavelet transform. Then, the Improved Deep Residual Network is built and the residual module structure is optimized. The Dropout layer is introduced and the global average pooling technique is used instead of the fully connected layer. Finally, the Improved Deep Residual Network is used as the pre-training network model and the tool wear state monitoring model combined with the model-based Transfer Learning method is constructed. The results show that the accuracy of the proposed monitoring method is up to 99.74%. The presented network model has the advantages of simple structure, small number of parameters, good robustness and reliability. The ideal classification effect can be achieved with fewer iterations.

View this article on IEEE Xplore