BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260525T171143EDT-48088A3TcV@132.216.98.100
DTSTAMP:20260525T211143Z
DESCRIPTION:Abstract\n\nMassive Multiple-Input Multiple-Output (M-MIMO) sys
 tems are fundamental to achieving the high data rates and reliability requ
 ired by future wireless networks. However\, realizing the full potential o
 f M-MIMO is hindered by significant challenges in symbol detection\, parti
 cularly the high computational complexity and performance degradation unde
 r realistic\, non-ideal operating conditions. Conventional detectors often
  struggle with complexity or rely on simplifying assumptions\, such as sta
 tionary channels\, white Gaussian noise\, and perfect channel state inform
 ation (CSI) knowledge\, which do not hold in practice. This thesis address
 es these critical limitations by developing and evaluating novel symbol de
 tection algorithms tailored for practical M-MIMO deployments.\n\nFirst\, w
 e introduce the Preconditioned Learned Conjugate Gradient Network (PrLcgNe
 t)\, a learning-based detector that accelerates training convergence in st
 ationary M-MIMO systems by incorporating a preconditioner during training.
  PrLcgNet achieves superior symbol error rate (SER) performance with reduc
 ed complexity compared to prior learning-based detectors. Building on this
 \, we extend PrLcgNet to Dynamic Conjugate Gradient Network (DyCoGNet)\, t
 ailored for time-varying channels. DyCoGNet leverages meta-learning and se
 lf-supervised learning guided by forward error correction (FEC)\, enabling
  rapid adaptation to unforeseen channel dynamics without labeled data\, ou
 tperforming conventional and recent self-supervised benchmarks.\n\nSecond\
 , we propose the Zero-Forcing based Latent Space Symbol Detector (ZF-LSSD)
  to effectively address the challenges associated with unknown or non-anal
 ytic noise distributions. ZF-LSSD combines zero-forcing initialization wit
 h score-based generative modeling using stochastic differential equations 
 (SDEs) and a localized search strategy. This approach facilitates efficien
 t approximate maximum likelihood detection within a latent space\, circumv
 enting the computational intractability posed by complex noise distributio
 ns in large-scale MIMO systems. Numerical simulations demonstrate that ZF-
 LSSD consistently outperforms existing benchmark methods across various ma
 ssive MIMO configurations and diverse additive noise scenarios.\n\nFinally
 \, we introduce Attention-Based Successive Interference Cancellation (ASIC
 )\, a novel detection method tailored for massive MIMO systems experiencin
 g imperfect CSI. ASIC integrates permutation-equivariant neural architectu
 res with CSI-derived priors\, dynamically adjusting the sequential decodin
 g process to mitigate residual interference and channel uncertainty. Disti
 nct from purely data-driven detectors\, ASIC strikes a balance between mod
 el-based inference and learned attention mechanisms\, significantly enhanc
 ing the robustness of SIC detectors under imperfect CSI conditions without
  incurring the substantial computational overhead typical of purely data-d
 riven methods. Simulation results demonstrate that ASIC consistently maint
 ains robustness across varying user counts and outperforms traditional SIC
  algorithms as well as learning-based baselines in scenarios with imperfec
 t CSI.\n\nCollectively\, this thesis contributes novel M-MIMO symbol detec
 tion techniques that exhibit enhanced robustness and adaptability to pract
 ical channel impairments\, thereby advancing the feasibility and performan
 ce of M-MIMO systems in real-world wireless communication scenarios.\n
DTSTART:20251008T130000Z
DTEND:20251008T150000Z
LOCATION:Room 603\, McConnell Engineering Building\, CA\, QC\, Montreal\, H
 3A 0E9\, 3480 rue University
SUMMARY:PhD defence of Toluwaleke Olutayo – Machine Learning Approaches to 
 Symbol Detection in Massive MIMO Wireless Systems
URL:/ece/channels/event/phd-defence-toluwaleke-olutayo
 -machine-learning-approaches-symbol-detection-massive-mimo-wireless-368100
END:VEVENT
END:VCALENDAR