Abstract: Recently, Deep Neural Networks (DNNs) have been used extensively for Automatic Modulation Classification (AMC). Due to their high complexity, DNNs are typically unsuitable for deployment at resource-constrained edge networks. They are also vulnerable to adversarial attacks, which is a significant security concern.

This seminar proposes a Rotated Binary Large ResNet (RBLResNet) for AMC, which can be deployed in the edge network due to its low complexity. The performance gap between the RBLResNet and existing architectures with floating-point weights and activations can be closed by two proposed ensemble methods: (i) Multilevel Classification (MC) and (ii) bagging multiple RBLResNets. The MC method achieves 93.39% accuracy at 10 dB across all 24 modulation classes in the Deepsig dataset. This performance is comparable to state-of-the-art, with 4.75 times lower memory and 1,214 times lower computation. Furthermore, RBLResNet demonstrates greater adversarial robustness than existing DNN models. The proposed MC method employing RBLResNets achieves a notable adversarial accuracy of 87.25% across a diverse range of Signal-to-Noise Ratios (SNRs), outperforming existing methods and well-established defense mechanisms, to the best of our knowledge. Low memory, low computation, and the highest adversarial robustness make it a better choice for robust AMC in low-power edge devices.

Event Details
Title: Binarized ResNet: Enabling Robust Automatic Modulation Classification at the resource-constrained Edge
Date: April 07, 2026 at 03:00 PM
Venue: Google Meet (https://meet.google.com/ajv-mymg-jps)
Speaker: Mr. Nitin Priyadarshinishankar (EE20D425)
Guide: Dr. Sheetal Kalyani
Type: PHD seminar

Updated: