Abstract: Dense IoT networks require reliable communication despite limited spectrum and substantial multi-user interference while maintaining manageable receiver complexity. We discuss a deep-learning-based end-to-end multi-user communication design for interference-limited finite-blocklength communication, focusing on short and medium blocklengths. We extend a prior 2-user SiameseNet transceiver framework to accommodate 2, 4, and 8 users, leveraging learned redundancy to suppress the interference and enhance the noise robustness. Compared to conventional non-orthogonal access baselines, our method demonstrates strong Block Error Rate (BLER) performance across various scenarios without resorting to joint detection; the per-user decoder scales roughly linearly with the number of users. Further, we examine the robustness under interference mismatch and unequal interference strengths, critical for practical deployments with heterogeneous devices. A study of the correlation analysis reveals that the learned codeword orthogonality increases with the increase in the interference between the user pairs, thereby corroborating the observed BLER improvements. In addition, we present preliminary results for a 2X2 MIMO setup under fixed-channel CSIT and CSIR, indicating potential to extend the framework to IoT gateways with multiple antennas.
Event Details
Title: Deep Learning-Based Multi-User Communication Design: Interference-Aware Finite-Block Length Communication
Date: June 16, 2026 at 10:00 AM
Venue: Google Meet (http://meet.google.com/qoi-pxiu-doh)
Speaker: Mr. Arkadeep Sinha (EE21S088)
Guide: Dr. Manivasakan R
Type: MS seminar

Updated: