Searcharxiv⌕ Search

arXiv subjects

Yaser Mike Banad

Publications and source records attributed to Yaser Mike Banad.

At least 19 recordsLinked to original sources
arXiv2026

Synthesizability Prediction of Crystalline Structures with Structure-Aware Feature Learning and Uncertainty Quantification

Predicting which hypothetical inorganic crystals can be experimentally realized remains a central challenge in accelerating materials discovery. SyntheFormer is a positive-unlabeled framework that learns synthesizability directly from crystal structure, combining Fourier-transformed crystal properties (FTCP) representation with structure-aware feature extraction, Random-Forest feature selection, and a compact deep MLP classifier. The model is trained on historical data from 2011 through 2018 and evaluated prospectively on future years from 2019 to 2025, where the positive class constitutes only 1.02 percent of samples. Under this temporally separated evaluation, SyntheFormer achieves a test AUC of 0.735, AUPRC of 0.099 and 97.6 percent recall at 94.2 percent coverage with dual-threshold calibration. Direct prospective validation supports this result as two materials that were unlabeled at the time of data curation, Y6Fe(SiS7)2 and BaYb2F8, were assigned high SyntheFormer scores (0.961 and 0.753, respectively) and were subsequently reported experimentally. Crucially, the model recovers experimentally confirmed metastable compounds that lie far from the convex hull and simultaneously assigns low scores to many thermodynamically stable yet unsynthesized candidates, demonstrating that stability alone is insufficient to predict experimental attainability.

cond-mat.mtrl-sci↗
arXiv2026

SparseVoxelDet: Fully Sparse Voxel Networks for Efficient Event-Based Drone Detection

Event cameras excel at detecting small, fast drones, but today's detectors give away their key advantage: they convert the sparse event stream into dense grids and pay dense-processing cost on inputs that are almost entirely empty. We present SparseVoxelDet, to our knowledge the first coordinate-sparse 3D event voxel bounding-box detector: backbone, feature pyramid, temporal reduction, and detection head all operate directly on coordinate-indexed features, with no dense spatial grid at any stage. Building it exposed a hidden failure mode we name support inflation: an input filling a median 0.0652% of the voxel lattice inflates stage by stage until standard pyramid fusion leaves the detection head locally near-dense. We answer with two ideas. Expansion-free inverse-convolution fusion provably creates no active sites beyond the stored backbone supports, cutting head occupancy from a median 78.88% to 10.53%; quality-aligned supervision then recovers more accuracy than preserving sparsity costs. The payoff is measured, not assumed: executing the same trained network densely costs a median 27.5x the work and 4.65x the latency across 5,000 paired frames, with no frame cheaper dense, while the 6.22M-parameter detector reaches 87.01 AP50 on the FRED drone benchmark, ahead of matched dense controls, and holds its lead on the held-out test partition evaluated once. Sparsity, preserved by construction and supervised well, delivers the efficiency and the accuracy together.

cs.CV↗
arXiv2026

Unified Complex-valued Neural Network: A Magnitude-Phase Computational Model for Event-Driven Neuromorphic Learning

Artificial neural networks (ANN) provide accurate continuous-valued representation, whereas spiking neural networks (SNN) offer event-driven temporal processing, yet both paradigms face limitations when value encoding and timing dynamics must be learned within a single computational structure. This paper introduces a network based on Unified Complex-valued Neuron (UCN), a new neural computational model that integrates continuous activation and phase-driven event generation through an asymmetric complex-valued state. In the UCN, magnitude encodes signal strength while phase governs intrinsic temporal evolution and valued spike emission. A foundational training framework combining backpropagation (BP) and backpropagation through time (BPTT) is first developed to optimize magnitude and phase pathways in a unified way. To reduce computational complexity, an event-driven adaptive phase learning (EAPL) rule is then introduced as a more efficient alternative. The proposed model is evaluated through object tracking and Lorenz attractor learning. Results demonstrate that UCN-based Network (UCNN) provides accurate, stable, and interpretable spatiotemporal learning while preserving sparse event-driven computation for neuromorphic and edge-AI applications.

cs.NE↗
arXiv2026

Physics-informed generative AI for semiconductor manufacturing: Enforcing hard physical constraints in generative models by construction

Generative models are increasingly used to propose designs, data, and control actions for physical systems, yet many such systems are governed by hard physical constraints rather than by perceptual plausibility. Semiconductor manufacturing provides a demanding test case: generated masks, layouts, synthetic defect data, and process recipes must obey lithography, transport, reaction, and device-physics constraints, because physically invalid samples are not merely low quality but unusable. This Perspective argues that semiconductor manufacturing exposes a broader computational-science challenge, namely that generative AI for constrained physical domains must be physics-informed by construction, not corrected only through post-hoc filtering. We survey the emerging architectural toolkit, including physics-informed diffusion, PDE-constrained variational models, neural-operator priors, and conservation-law-respecting generative networks, and show how it connects to differentiable lithography, TCAD, process simulation, and autonomous experimentation. We identify four integration patterns between generative models and physics-based simulators, and we propose a research agenda centered on physics-fidelity benchmarks, differentiable simulator infrastructure, and multimodal foundation models for physical design and manufacturing. The central claim is analytical rather than rhetorical: where physical validity is the binding criterion of success, architectures that enforce it by construction should be expected to outperform those that filter for it after the fact, and the fab is the setting where this distinction is sharpest.

cs.LG↗
arXiv2026

Brain-inspired spike-timing plasticity for reliable label-efficient event-camera vision

Deploying event-camera object detectors is constrained by per-frame labeling requirements and GPU compute demands. This work introduces three local spike-timing-dependent plasticity (STDP) modules, including sequence, candidate, and tube-reliability modules, that operate on a single CPU thread without GPU support. On the FRED drone benchmark, the proposed framework spans three label-efficient supervision tiers. A strict zero-label detector achieves 53.8% mAP@30, approximately 26 train-derived bits achieve 76.9% mAP@30, and an STDP candidate-reliability gate achieves 78.60 +/- 0.42% mAP@30. Under acquisition-order drift, the cohort gate outperforms streaming k-means by 2.03 +/- 0.58 percentage points across 20 of 20 positive trials, while a no-drift control falsifies the effect. STDP reduces single-model variance by 6.6 times, and one trained gate matches a 44-seed ensemble bound. The gate transfers to Intel Lava with 89% top-2 agreement. On the EVUAV benchmark, a tube-level STDP layer reduces false alarms from 454 to 331e-4 at Pd >= 88%. Dense gradient-trained detectors cannot provide this combination of gradient training, dense matrix multiplication, and local plasticity-free operation by construction.

cs.CV↗
arXiv2026

AM-DefectNet: Additive Manufacturing Defect Classification Using Machine Learning -- A comparative Study

Additive Manufacturing (AM) processes present challenges in monitoring and controlling material properties and process parameters, affecting production quality and defect detection. Machine Learning (ML) techniques offer a promising solution for addressing these challenges. In this study, we introduce a comprehensive framework, AM-DefectNet, for benchmarking ML models in melt pool characterization, a critical aspect of AM. We evaluate 15 ML models across 10 metrics using 1514 training and 505 test datasets. Our benchmarking reveals that non-linear tree-based algorithms, particularly CatBoost, LGBM, and XGBoost, outperform other models, achieving accuracies of 92.47%, 91.08%, and 90.89%, respectively. Notably, the Deep Neural Network (DNN) also demonstrates competitive performance with an accuracy of 88.55%. CatBoost emerges as the top-performing algorithm, exhibiting superior performance in precision, recall, F1-score, and overall accuracy for defect classification tasks. Learning curves provide insights into model performance and data requirements, indicating potential areas for improvement. Our study highlights the effectiveness of ML models in melt pool characterization and defect detection, laying the groundwork for process optimization in AM.

cond-mat.mes-hall↗
arXiv2026

Multi-objective Evolutionary Algorithms (MOEAs) in PMEDM -- A Comparative Study in Pareto Frontier

Electrical discharge machining (EDM) is a crucial process in precision manufacturing, leveraging electro-thermal energy to remove material without electrode contact. In this study, we delve into the realm of Machine Learning (ML) to enhance the efficiency and precision of EDM, particularly focusing on Powder-Mixed Electrical Discharge Machining (PMEDM) with the integration of a vibration system. We comprehensively evaluate four leading ML models - Deep Neural Network (DNN), Extreme Gradient Boosting (XGBoost), Adaptive Gradient Boosting (AdaBoost), and ElasticNet, against a pool of ML models, employing various evaluation metrics including Accuracy, Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). Our evaluations, conducted on datasets enriched with features derived from powder addition and electrode vibration, reveal XGBoost superior accuracy, followed by AdaBoost, DNN, and ElasticNet. Furthermore, through the integration of Multi-Objective Evolutionary Algorithms (MOEAs) such as NSGA-II, NSGA-III, UNSGA-III, and C-TAEA, we explore and optimize the Pareto front to attain optimal solutions. Our findings underscore the transformative potential of ML and optimization techniques in advancing EDM processes, offering cost-effective, time-efficient, and reliable solutions for precision manufacturing applications.

cond-mat.mes-hall↗
arXiv2026

Memristor Technologies for Dynamic Vision Sensors: A Critical Assessment and Research Roadmap

Edge-AI deployment is bottlenecked by data-movement energy; pairing event-driven vision sensors with in-memory analog compute could lift that ceiling by orders of magnitude. Both technologies are individually mature; the framework distinguishing fabricated demonstrations from projected systems is missing. Of six application domains surveyed (robotics, autonomous vehicles, AR/VR, surveillance, medical imaging, IoT), half rest entirely on projection, and existing hardware sits at Technology Readiness Levels 2-5. This evidence-graded review applies a three-paradigm architectural taxonomy and benchmarks the gap against current digital neuromorphic alternatives. It identifies an end-to-end integrated DVS-memristor system as the field's open challenge, with testable accuracy and power targets.

cs.AR↗
arXiv2026

IoT-Enhanced CNN-Based Labelled Crack Detection for Additive Manufacturing Image Annotation in Industry 4.0

This paper presents an IoT-enhanced deep learning framework for automated crack detection in Additive Manufacturing (AM) surfaces using convolutional neural networks (CNNs). By integrating IoT-enabled real-time monitoring, high-resolution imaging, and edge computing, the system enables continuous in-situ defect detection and classification. Real-time data acquisition supports immediate CNN-based analysis, improving both accuracy and efficiency in AM quality control. The framework supports supervised and semi-supervised learning, enabling robust performance on large, sparsely annotated datasets. Using LabelImg for annotation and OpenCV for preprocessing, the system achieves 99.54% accuracy on 14,982 images, with 96% precision, 98% recall, and a 97% F1-score. Dataset balancing and augmentation significantly improve generalization, increasing accuracy from 32% to 99%. Beyond detection, the framework establishes a linkage between AM process parameters, defect formation, and surface topology, supporting predictive analytics and defect mitigation. Aligned with Industry 4.0, it incorporates Digital Twin (DT) technology for real-time process simulation, predictive maintenance, and adaptive control. Key contributions include an IoT-based monitoring system using edge devices (Raspberry Pi 4B), an optimized CNN with model quantization and batch processing reducing inference latency by 47%, and an MQTT-based low-latency data streaming system with 5G connectivity, lowering transmission overhead by 35%. DT integration further enables predictive defect analysis and dynamic adjustment of AM parameters. This work advances intelligent AM quality control by providing a scalable, high-accuracy, and low-latency framework. Future directions include multimodal data fusion, hybrid architectures, and enhanced Digital Twin simulations for AI-driven defect prevention.

cs.CV↗
arXiv2026

A Unified Phase-native Computational Principle Governs Hippocampal Spike Timing and Neural Coding

Hippocampal neurons exhibit precise phase locking to network oscillations, but the computational principle governing this temporal precision is still unclear. Neural information is conveyed jointly by firing rates and spike timing, but existing models treat these dimensions separately, limiting mechanistic interpretation of spike-field coupling and its reported association with spectral features such as the aperiodic slope. Here we show that hippocampal phase locking emerges from a fundamental dynamical mechanism referred to as forced phase integration that separates neural information into orthogonal magnitude (what) and phase (when) coordinates. To formalize this principle, the unified complex-valued neuron (UCN) has been developed, a biologically grounded generative framework in which spike timing arises from phase accumulation while spike magnitude encodes instantaneous signal strength. This framework reproduces biological spike-theta synchronization and enables mechanistic re-evaluation of slope-locking associations, demonstrating that previously reported effects arise from oscillatory contamination rather than causal modulation. These findings establish a unified phase-native principle of neural timing and coding.

q-bio.NC↗
arXiv2026

Toward Digital Twins in 3D IC Packaging: A Critical Review of Physics, Data, and Hybrid Architectures

Three-dimensional integrated circuit (3D IC) pack-aging and heterogeneous integration have emerged as central pillars of contemporary semiconductor scaling. Yet, the multi-physics coupling inherent to stacked architectures manifesting as thermal hot spots, warpage-induced stresses, and interconnect aging demands monitoring and control capabilities that surpass traditional offline metrology. Although Digital Twin (DT) technology provides a principled route to real-time reliability management, the existing literature remains fragmented and frequently blurs the distinction between static multiphysics simulation workflows and truly dynamic, closed-loop twins. This critical review distinguishes itself by addressing these deficiencies through three specific contributions. First, we clarify the Digital Twin hierarchy to resolve terminological ambiguity between digital models, shadows, and twins. Second, we synthesize three foundational enabling technologies: (1) physics-based modeling, emphasizing the shift from computationally intensive finite-element analysis (FEA) to real-time surrogate models; (2) data-driven paradigms, highlighting virtual metrology (VM) for inferring latent metrics; and (3) in-situ sensing, the nervous system coupling the physical stack to its virtual counterpart. Third, beyond a descriptive survey, we propose a unified hybrid DT architecture that leverages physics-informed machine learning (e.g., PINNs) to reconcile data scarcity with latency constraints. Finally, we outline a standards-aligned roadmap incorporating IEEE 1451 and UCIe protocols to accelerate the transition from passive digital shadows to autonomous, self-optimizing Digital Twins for 3D IC manufacturing and field operation.

cs.AR↗
arXiv2026

Event-based Heterogeneous Information Processing for Online Vision-based Obstacle Detection and Localization

This paper introduces a novel framework for robotic vision-based navigation that integrates Hybrid Neural Networks (HNNs) with Spiking Neural Network (SNN)-based filtering to enhance situational awareness for unmodeled obstacle detection and localization. By leveraging the complementary strengths of Artificial Neural Networks (ANNs) and SNNs, the system achieves both accurate environmental understanding and fast, energy-efficient processing. The proposed architecture employs a dual-pathway approach: an ANN component processes static spatial features at low frequency, while an SNN component handles dynamic, event-based sensor data in real time. Unlike conventional hybrid architectures that rely on domain conversion mechanisms, our system incorporates a pre-developed SNN-based filter that directly utilizes spike-encoded inputs for localization and state estimation. Detected anomalies are validated using contextual information from the ANN pathway and continuously tracked to support anticipatory navigation strategies. Simulation results demonstrate that the proposed method offers acceptable detection accuracy while maintaining computational efficiency close to SNN-only implementations, which operate at a fraction of the resource cost. This framework represents a significant advancement in neuromorphic navigation systems for robots operating in unpredictable and dynamic environments.

cs.RO↗
arXiv2026

Machine learning model for predicting surface wettability in laser-textured metal alloys

Surface wettability, governed by both topography and chemistry, plays a critical role in applications such as heat transfer, lubrication, microfluidics, and surface coatings. In this study, we present a machine learning (ML) framework capable of accurately predicting the wettability of laser-textured metal alloys using experimentally derived morphological and chemical features. Superhydrophilic and superhydrophobic surfaces were fabricated on AA6061 and AISI 4130 alloys via nanosecond laser texturing followed by chemical immersion treatments. Surface morphology was quantified using the Laws texture energy method and profilometry, while surface chemistry was characterized through X-ray photoelectron spectroscopy (XPS), extracting features such as functional group polarity, molecular volume, and peak area fraction. These features were used to train an ensemble neural network model incorporating residual connections, batch normalization, and dropout regularization. The model achieved high predictive accuracy (R2 = 0.942, RMSE = 13.896), outperforming previous approaches. Feature importance analysis revealed that surface chemistry had the strongest influence on contact angle prediction, with topographical features also contributing significantly. This work demonstrates the potential of artificial intelligence to model and predict wetting behavior by capturing the complex interplay of surface characteristics, offering a data-driven pathway for designing tailored functional surfaces.

cs.LG↗
arXiv2025

XG-Attention-WGAN PIC: Utilizing XGboost-Attention-WGAN for Photonics Integrated Circuit Design

Photonic Integrated Circuits (PICs) are fundamental for optical computing, communication, quantum information processing, and precision sensing. However, traditional numerical simulations for designing PIC components are computationally intensive and struggle with high-dimensional parameter spaces. This paper introduces XG-Attention-WGAN PIC, a novel framework that synergistically combines Wasserstein Generative Adversarial Networks (WGANs) with eXtreme Gradient Boosting (XGBoost) to enhance parameter prediction and inverse design in PICs. We utilize Finite-Difference Time-Domain simulations to generate high-fidelity training data, which is augmented by WGAN-generated synthetic data, yielding a root mean squared error (RMSE) of 0.26089. When integrated with XGBoost, this error is reduced to 0.008. The integration of a 64-head self-attention mechanism within the WGAN generator significantly improves data quality and model efficiency over 1000 training epochs. Demonstrated on microring resonators, our approach not only achieves superior prediction accuracy and design optimization but also autonomously discovers a novel, experimentally realizable geometry with enhanced Q-factor performance. The proposed framework provides a scalable, data-driven strategy for developing high-performance PIC components, with promising implications for quantum computing and advanced optical systems.

physics.optics↗
arXiv2025

LabelImg: CNN-Based Surface Defect Detection

In the journey of computer vision system development, the acquisition and utilization of annotated images play a central role, providing information about object identity, spatial extent, and viewpoint in depicted scenes. However, thermal manufacturing processes like Laser Powder Bed Fusion (LPBF) often yield surfaces with defects such as Spatter, Crack, Pinhole, and Hole due to the Balling phenomenon. Preprocessing images from LPBF, riddled with defects, presents challenges in training machine learning (ML) algorithms. Detecting defects is critical for predicting production quality and identifying crucial points in artificial or natural structures. This paper introduces a deep learning-based approach utilizing Convolutional Neural Networks (CNNs) to automatically detect and segment surface defects like cracks, spatter, holes, and pinholes on production surfaces. In contrast to traditional machine learning techniques requiring extensive processing time and manual feature crafting, deep learning proves more accurate. The proposed architecture undergoes training and testing on 14,982 labeled images annotated using the LabelImg tool. Each object in the images is manually annotated with bounding boxes and segmented masks. The trained CNN, coupled with OpenCV preprocessing techniques, achieves an impressive 99.54% accuracy on the dataset with resolutions of 1536 x 1103 pixels. Evaluation metrics for 50 true crack tests demonstrate precision, recall, and F1-score exceeding 96%, 98%, and 97%, respectively. Similarly, for 124 true pinhole tests, the metrics are 99%, 100%, and 100%, for 258 true hole tests, they are 99%, 99%, and 99%, and for 318 spatter tests, the metrics are 100%, 99%, and 100%. These results highlight the precision and effectiveness of the entire process, showcasing its potential for reliable defect detection in production surfaces.

cond-mat.mes-hall↗
arXiv2025

TransMatch: A Transfer-Learning Framework for Defect Detection in Laser Powder Bed Fusion Additive Manufacturing

Surface defects in Laser Powder Bed Fusion (LPBF) pose significant risks to the structural integrity of additively manufactured components. This paper introduces TransMatch, a novel framework that merges transfer learning and semi-supervised few-shot learning to address the scarcity of labeled AM defect data. By effectively leveraging both labeled and unlabeled novel-class images, TransMatch circumvents the limitations of previous meta-learning approaches. Experimental evaluations on a Surface Defects dataset of 8,284 images demonstrate the efficacy of TransMatch, achieving 98.91% accuracy with minimal loss, alongside high precision, recall, and F1-scores for multiple defect classes. These findings underscore its robustness in accurately identifying diverse defects, such as cracks, pinholes, holes, and spatter. TransMatch thus represents a significant leap forward in additive manufacturing defect detection, offering a practical and scalable solution for quality assurance and reliability across a wide range of industrial applications.

cs.CV↗
arXiv2025

Novel Pigeon-inspired 3D Obstacle Detection and Avoidance Maneuver for Multi-UAV Systems

Recent advances in multi-agent systems manipulation have demonstrated a rising demand for the implementation of multi-UAV systems in urban areas, which are always subjected to the presence of static and dynamic obstacles. Inspired by the collective behavior of tilapia fish and pigeons, the focus of the presented research is on the introduction of a nature-inspired collision-free formation control for a multi-UAV system, considering the obstacle avoidance maneuvers. The developed framework in this study utilizes a semi-distributed control approach, in which, based on a probabilistic Lloyd's algorithm, a centralized guidance algorithm works for optimal positioning of the UAVs, while a distributed control approach has been used for the intervehicle collision and obstacle avoidance. Further, the presented framework has been extended to the 3D space with a novel definition of 3D maneuvers. Finally, the presented framework has been applied to multi-UAV systems in 2D and 3D scenarios, and the obtained results demonstrated the validity of the presented method in dynamic environments with stationary and moving obstacles.

cs.RO↗
arXiv2025

Neuromorphic Digital-Twin-based Controller for Indoor Multi-UAV Systems Deployment

Presented study introduces a novel distributed cloud-edge framework for autonomous multi-UAV systems that combines the computational efficiency of neuromorphic computing with nature-inspired control strategies. The proposed architecture equips each UAV with an individual Spiking Neural Network (SNN) that learns to reproduce optimal control signals generated by a cloud-based controller, enabling robust operation even during communication interruptions. By integrating spike coding with nature-inspired control principles inspired by Tilapia fish territorial behavior, our system achieves sophisticated formation control and obstacle avoidance in complex urban environments. The distributed architecture leverages cloud computing for complex calculations while maintaining local autonomy through edge-based SNNs, significantly reducing energy consumption and computational overhead compared to traditional centralized approaches. Our framework addresses critical limitations of conventional methods, including the dependency on pre-modeled environments, computational intensity of traditional methods, and local minima issues in potential field approaches. Simulation results demonstrate the system's effectiveness across two different scenarios. First, the indoor deployment of a multi-UAV system made-up of 15 UAVs. Then the collision-free formation control of a moving UAV flock including 6 UAVs considering the obstacle avoidance. Owing to the sparsity of spiking patterns, and the event-based nature of SNNs in average for the whole group of UAVs, the framework achieves almost 90% reduction in computational burden compared to traditional von Neumann architectures implementing traditional artificial neural networks.

cs.NE↗