arXiv subjects

Firas Darwish

Publications and source records attributed to Firas Darwish.

4 recordsLinked to original sources

arXiv2026

Leveraging Agonistic-Antagonistic Coactivation in Single-Grid HDsEMG for Hand Gesture Recognition

Surface Electromyography (sEMG) is critical for intention prediction in human-computer interfaces, such as for prosthetics control. Although deep learning models for Hand Gesture Recognition (HGR) yield excellent results, they impose high computational and hardware demands. This paper addresses this bottleneck by exploiting redundancies in agonist-antagonist muscle activity, hypothesizing that coactivations present in the sEMG signals from the extensor or flexor groups alone are sufficient for accurate HGR. We evaluate this by comparing convolutional neural networks (CNNs) trained on one muscle grid against CNN architectures trained jointly on both grids. Experiments were conducted using 16 gestures from a dataset of high-density sEMG signals from 20 subjects. The results demonstrate that the extensor grid alone achieves performance (89.5% balanced accuracy, 0.99 AUROC) comparable to the dual-grid system (94.6% balanced accuracy, 1.00 AUROC). Notably, even when applying slow joint fusion to capture spatial features across grids, model performance did not improve. GradCAM visualizations and anatomical analysis further indicate that the extensor region provides superior signal quality compared to the flexors. Our findings suggest that for a base set of DoF gestures, HGR hardware requirements and computational complexity can be halved without a prohibitive loss in accuracy.

eess.SP↗

arXiv2026

Finding the Time to Think: Learning Planning Budgets in Real-Time RL

Deliberating takes time. In real-time settings, that time is not free. Standard reinforcement learning (RL) sidesteps this as the environment waits indefinitely for the agent's decision. Instead, we study real-time RL environments where the environment progresses while waiting for the agent's action. Building on prior real-time formalizations, we introduce variable-delay real-time RL, where the agent chooses how long to deliberate at each decision point since the environment progresses. For the planning agents we use, the right delay is state-dependent, and naively planning how long to plan can paralyze the agent. We instead approach this setting by training a lightweight gating policy on top of a planner to select state-dependent planning budgets. Across real-time Pac-Man, Tetris, Snake, Speed Hex, and Speed Go, our gating policy outperforms fixed-budget and heuristic baselines, and transfers to a real-time setup where the environment and agent run on two different GPUs.

cs.LG↗

arXiv2026

Motion Capture is Not the Target Domain: Scaling Synthetic Data for Learning Motion Representations

Synthetic data offers a compelling path to scalable pretraining when real-world data is scarce, but models pretrained on synthetic data often fail to transfer reliably to deployment settings. We study this problem in full-body human motion, where large-scale data collection is infeasible but essential for wearable-based Human Activity Recognition (HAR), and where synthetic motion can be generated from motion-capture-derived representations. We pretrain motion time-series models using such synthetic data and evaluate their transfer across diverse downstream HAR tasks. Our results show that synthetic pretraining improves generalisation when mixed with real data or scaled sufficiently. We also demonstrate that large-scale motion-capture pretraining yields only marginal gains due to domain mismatch with wearable signals, clarifying key sim-to-real challenges and the limits and opportunities of synthetic motion data for transferable HAR representations.

cs.LG↗

arXiv2025

Uncertainty Quantification for Machine Learning in Healthcare: A Survey

Uncertainty Quantification (UQ) is pivotal in enhancing the robustness, reliability, and interpretability of Machine Learning (ML) systems for healthcare, optimizing resources and improving patient care. Despite the emergence of ML-based clinical decision support tools, the lack of principled quantification of uncertainty in ML models remains a major challenge. Current reviews have a narrow focus on analyzing the state-of-the-art UQ in specific healthcare domains without systematically evaluating method efficacy across different stages of model development, and despite a growing body of research, its implementation in healthcare applications remains limited. Therefore, in this survey, we provide a comprehensive analysis of current UQ in healthcare, offering an informed framework that highlights how different methods can be integrated into each stage of the ML pipeline including data processing, training and evaluation. We also highlight the most popular methods used in healthcare and novel approaches from other domains that hold potential for future adoption in the medical context. We expect this study will provide a clear overview of the challenges and opportunities of implementing UQ in the ML pipeline for healthcare, guiding researchers and practitioners in selecting suitable techniques to enhance the reliability, safety and trust from patients and clinicians on ML-driven healthcare solutions.

cs.LG↗