Omid Yazdani | Searcharxiv

arXiv2026

Beam Squint Mitigation in Wideband Hybrid Beamformers: Full-TTD, Sparse-TTD, or Non-TTD?

Beam squint poses a fundamental challenge in wideband hybrid beamforming, particularly for mmWave and THz systems that demand both ultra-wide bandwidth and high directional beams. While conventional phase shifter-based beamformers may offer partial mitigation, True Time Delay (TTD) units provide a fundamentally more effective solution by enabling frequency-independent beam steering. However, the high cost of TTD units has recently driven much interest in Sparse-TTD architectures, which combine a limited number of TTDs with a higher number of conventional PSs to balance performance and cost. This paper provides a critical examination of beam squint mitigation strategies in wideband hybrid beamformers, comparing Full-TTD, Sparse-TTD, and Non-TTD architectures. We analyze recent Non-TTD approaches, specifically the scheme leveraging the wideband beam gain (WBBG) concept, evaluating their performance and cost characteristics against TTD-based solutions. A key focus is placed on the practical limitations of Sparse-TTD architectures, particularly the often-overlooked requirement for wideband PSs operating alongside TTDs, which can significantly impact performance and implementation cost in real-world scenarios, especially for ultra-wideband applications. Finally, we conduct a cost-performance analysis to examine the trade-offs inherent in each architecture and provide guidance on selecting the most suitable hybrid beamforming structure for various fractional bandwidth regimes.

eess.SY↗

arXiv2025

Large Language Models versus Classical Machine Learning: Performance in COVID-19 Mortality Prediction Using High-Dimensional Tabular Data

This study compared the performance of classical feature-based machine learning models (CMLs) and large language models (LLMs) in predicting COVID-19 mortality using high-dimensional tabular data from 9,134 patients across four hospitals. Seven CML models, including XGBoost and random forest (RF), were evaluated alongside eight LLMs, such as GPT-4 and Mistral-7b, which performed zero-shot classification on text-converted structured data. Additionally, Mistral- 7b was fine-tuned using the QLoRA approach. XGBoost and RF demonstrated superior performance among CMLs, achieving F1 scores of 0.87 and 0.83 for internal and external validation, respectively. GPT-4 led the LLM category with an F1 score of 0.43, while fine-tuning Mistral-7b significantly improved its recall from 1% to 79%, yielding a stable F1 score of 0.74 during external validation. Although LLMs showed moderate performance in zero-shot classification, fine-tuning substantially enhanced their effectiveness, potentially bridging the gap with CML models. However, CMLs still outperformed LLMs in handling high-dimensional tabular data tasks. This study highlights the potential of both CMLs and fine-tuned LLMs in medical predictive modeling, while emphasizing the current superiority of CMLs for structured data analysis.

cs.LG↗