Abstract--We introduce Iwin Transformer, a novel position-embedding-free hierarchical vision transformer, which can be fine-tuned directly from low to high resolution, through the collaboration of innovative interleaved window attention and depthwise separable convolution. This approach uses attention to connect distant tokens and applies convolution to link neighboring tokens, enabling global information exchange within a single module, overcoming Swin Transformer's limitation of requiring two consecutive blocks to approximate global attention. Extensive experiments on visual benchmarks demonstrate that Iwin Transformer exhibits strong competitiveness in tasks such as image classification (87.4 top-1 accuracy on ImageNet-1K), semantic segmentation and video action recognition. We also validate the effectiveness of the core component in Iwin as a standalone module that can seamlessly replace the self-attention module in class-conditional image generation. The concepts and methods introduced by the Iwin Transformer have the potential to inspire future research, like Iwin 3D Attention in video generation. ISION Transformers (ViTs) [1] have have fundamentally transformed computer vision by borrowing the transformer architecture from natural language models [2]. Unlike Convolutional Neural Networks (CNNs) [3], which rely on local receptive fields to capture image features, ViTs leverage self-attention mechanisms to get global dependencies, demonstrating remarkable performance on vision tasks. To tackle the challenge of quadratic complexity in Vision Transformers (ViTs) and enhance their efficiency while maintaining performance, various approaches have been proposed. Hierarchical Designs such as PVT [4] and Twins [5] utilize multi-scale feature pyramids to progressively reduce spatial dimensions. Hybrid CNN-Transformer Architectures like Con-ViT [6] and CoAtNet [7] combine convolutional operations with self-attention to leverage the strengths of both paradigms. Efficient Token Fusion strategies such as TokenLearner [8] dynamically aggregate tokens to reduce sequence length, while Sparse Attention Patterns exemplified by Reformer [9] utilize locality-sensitive hashing to attend only to relevant tokens. Additionally, efficient implementations like Performer [10] approximate attention through kernel methods to achieve linear complexity.

Abstract--The increasing demand for on-device intelligence in Edge AI and TinyML applications requires the efficient execution of modern Convolutional Neural Networks (CNNs). While lightweight architectures like MobileNetV2 employ Depth-wise Separable Convolutions (DSC) to reduce computational complexity, their multi-stage design introduces a critical performance bottleneck inherent to layer-by-layer execution: the high energy and latency cost of transferring intermediate feature maps to either large on-chip buffers or off-chip DRAM. T o address this memory wall, this paper introduces a novel hardware accelerator architecture that utilizes a fused pixel-wise dataflow. Implemented as a Custom Function Unit (CFU) for a RISC-V processor, our architecture eliminates the need for intermediate buffers entirely, reducing the data movement up to 87% compared to conventional layer-by-layer execution. It computes a single output pixel to completion across all DSC stages-expansion, depthwise convolution, and projection-by streaming data through a tightly-coupled pipeline without writing to memory. Evaluated on a Xilinx Artix-7 FPGA, our design achieves a speedup of up to 59.3x over the baseline software execution on the RISC-V core. Furthermore, ASIC synthesis projects a compact 0.284 mm This work confirms the feasibility of a zero-buffer dataflow within a TinyML resource envelope, offering a novel and effective strategy for overcoming the memory wall in edge AI accelerators. Edge AI[1] involves running artificial intelligence algorithms directly on local hardware, such as sensors and Internet of Things (IoT) units, bringing computation to the source of data creation. This allows for real-time processing without constant reliance on the cloud, an approach that offers several key benefits: low latency due to local processing, enhanced privacy by keeping sensitive data on the device, and reduced network bandwidth consumption, which enables reliable of-fline operation.[2] A critical subfield of this domain is Tiny Machine Learning (TinyML)[3], which specifically focuses on deploying machine learning models directly onto low-cost, ultra-low-power microcontrollers (MCUs) and embedded systems. These devices operate under severe constraints, often with power budgets in the milliwatt range and with only a few hundred kilobytes of memory, making on-device intelligence a significant technical challenge. The typical TinyML workflow involves taking a fully trained model and optimizing it for on-device inference by applying techniques such as quantization and pruning to create a smaller, more efficient model in a compact format.

Direction-of-Arrival (DOA) estimation is critical in spatial audio and acoustic signal processing, with wide-ranging applications in real-world. Most existing DOA models are trained on synthetic data by convolving clean speech with room impulse responses (RIRs), which limits their generalizability due to constrained acoustic diversity. In this paper, we revisit DOA estimation using a recently introduced dataset constructed with the assistance of large language models (LLMs), which provides more realistic and diverse spatial audio scenes. We benchmark several representative neural-based DOA methods on this dataset and propose LightDOA, a lightweight DOA estimation model based on depthwise separable convolutions, specifically designed for mutil-channel input in varying environments. Experimental results show that LightDOA achieves satisfactory accuracy and robustness across various acoustic scenes while maintaining low computational complexity. This study not only highlights the potential of spatial audio synthesized with the assistance of LLMs in advancing robust and efficient DOA estimation research, but also highlights LightDOA as efficient solution for resource-constrained applications.

Autism Spectrum Disorder significantly influences the communication abilities, learning processes, behavior, and social interactions of individuals. Although early intervention and customized educational strategies are critical to improving outcomes, there is a pivotal gap in understanding and addressing nuanced behavioral patterns and emotional identification in autistic children prior to skill development. This extended research delves into the foundational step of recognizing and mapping these patterns as a prerequisite to improving learning and soft skills. Using a longitudinal approach to monitor emotions and behaviors, this study aims to establish a baseline understanding of the unique needs and challenges faced by autistic students, particularly in the Information Technology domain, where opportunities are markedly limited. Through a detailed analysis of behavioral trends over time, we propose a targeted framework for developing applications and technical aids designed to meet these identified needs. Our research underscores the importance of a sequential and evidence-based intervention approach that prioritizes a deep understanding of each child's behavioral and emotional landscape as the basis for effective skill development. By shifting the focus toward early identification of behavioral patterns, we aim to foster a more inclusive and supportive learning environment that can significantly improve the educational and developmental trajectory of children with ASD.

B. Observation History The core contribution of this work is a novel history encoder for processing historical observations, which integrates two key operations: depthwise separable convolution and multi-head attention. The background of these operations is briefly reviewed below. Depthwise separable convolution [33] is a streamlined variant of standard convolution that reduces both parameter count and computational cost. It decomposes the operation into two steps: (1) a depthwise convolution, which applies a single filter to each input channel, and (2) a pointwise convolution, which uses a 1 1 convolution to linearly combine the outputs of the depthwise stage. This factorization enables efficient extraction of spatial and cross-channel features while maintaining strong representational capacity. It has been widely adopted in lightweight neural architectures such as MobileNet [34] and is particularly well suited to real-time and resource-constrained applications. Multi-head attention [9] is a fundamental component of Transformer architectures, enabling the model to capture diverse patterns across different representation subspaces.

Bioprinting is a rapidly advancing field that offers a transformative approach to fabricating tissue and organ models through the precise deposition of cell-laden bioinks. Ensuring the fidelity and consistency of printed structures in real-time remains a core challenge, particularly under constraints imposed by limited imaging data and resource-constrained embedded hardware. Semantic segmentation of the extrusion process, differentiating between nozzle, extruded bioink, and surrounding background, enables in situ monitoring critical to maintaining print quality and biological viability. In this work, we introduce a lightweight semantic segmentation framework tailored for real-time bioprinting applications. We present a novel, manually annotated dataset comprising 787 RGB images captured during the bioprinting process, labeled across three classes: nozzle, bioink, and background. To achieve fast and efficient inference suitable for integration with bioprinting systems, we propose a BioLite U-Net architecture that leverages depthwise separable convolutions to drastically reduce computational load without compromising accuracy. Our model is benchmarked against MobileNetV2 and MobileNetV3-based segmentation baselines using mean Intersection over Union (mIoU), Dice score, and pixel accuracy. All models were evaluated on a Raspberry Pi 4B to assess real-world feasibility. The proposed BioLite U-Net achieves an mIoU of 92.85% and a Dice score of 96.17%, while being over 1300x smaller than MobileNetV2-DeepLabV3+. On-device inference takes 335 ms per frame, demonstrating near real-time capability. Compared to MobileNet baselines, BioLite U-Net offers a superior tradeoff between segmentation accuracy, efficiency, and deployability, making it highly suitable for intelligent, closed-loop bioprinting systems.

The "Repeat" represents the maximum number of repeated blocks in a group. " represents the inverted bottleneck " represents the inverted bottleneck

Automated structural defect segmentation in civil infrastructure faces a critical challenge: achieving high accuracy while maintaining computational efficiency for real-time deployment. This paper presents FORTRESS (Function-composition Optimized Real-Time Resilient Structural Segmentation), a new architecture that balances accuracy and speed by using a special method that combines depthwise separable convolutions with adaptive Kolmogorov-Arnold Network integration. FORTRESS incorporates three key innovations: a systematic depthwise separable convolution framework achieving a 3.6x parameter reduction per layer, adaptive TiKAN integration that selectively applies function composition transformations only when computationally beneficial, and multi-scale attention fusion combining spatial, channel, and KAN-enhanced features across decoder levels. The architecture achieves remarkable efficiency gains with 91% parameter reduction (31M to 2.9M), 91% computational complexity reduction (13.7 to 1.17 GFLOPs), and 3x inference speed improvement while delivering superior segmentation performance. Evaluation on benchmark infrastructure datasets demonstrates state-of-the-art results with an F1- score of 0.771 and a mean IoU of 0.677, significantly outperforming existing methods including U-Net, SA-UNet, and U- KAN. The dual optimization strategy proves essential for optimal performance, establishing FORTRESS as a robust solution for practical structural defect segmentation in resource-constrained environments where both accuracy and computational efficiency are paramount. Comprehensive architectural specifications are provided in the Supplemental Material. Source code is available at URL: https://github.com/faeyelab/fortress-paper-code.

Cardiac arrhythmias are a leading cause of life-threatening cardiac events, highlighting the urgent need for accurate and timely detection. Electrocardiography (ECG) remains the clinical gold standard for arrhythmia diagnosis; however, manual interpretation is time-consuming, dependent on clinical expertise, and prone to human error. Although deep learning has advanced automated ECG analysis, many existing models abstract away the signal's intrinsic temporal and morphological features, lack interpretability, and are computationally intensive-hindering their deployment on resource-constrained platforms. In this work, we propose two novel lightweight 1D convolutional neural networks, ArrhythmiNet V1 and V2, optimized for efficient, real-time arrhythmia classification on edge devices. Inspired by MobileNet's depthwise separable convolutional design, these models maintain memory footprints of just 302.18 KB and 157.76 KB, respectively, while achieving classification accuracies of 0.99 (V1) and 0.98 (V2) on the MIT-BIH Arrhythmia Dataset across five classes: Normal Sinus Rhythm, Left Bundle Branch Block, Right Bundle Branch Block, Atrial Premature Contraction, and Premature Ventricular Contraction. In order to ensure clinical transparency and relevance, we integrate Shapley Additive Explanations and Gradient-weighted Class Activation Mapping, enabling both local and global interpretability. These techniques highlight physiologically meaningful patterns such as the QRS complex and T-wave that contribute to the model's predictions. We also discuss performance-efficiency trade-offs and address current limitations related to dataset diversity and generalizability. Overall, our findings demonstrate the feasibility of combining interpretability, predictive accuracy, and computational efficiency in practical, wearable, and embedded ECG monitoring systems.