How Much Does It Cost to Answer My Question? Benchmarking Cloud VLM-based VQA Systems

Vision-language models (VLMs) are becoming a practical backend for mobile visual question answering (VQA) systems, enabling smartphones and smart glasses to answer users' questions about the physical world. Since modern VLMs remain difficult to run on mobile and edge devices, VQA systems increasingly offload inference to cloud-based VLMs. This gives mobile devices access to stronger computation, but it also makes visual input preparation a key system variable: how the image is prepared before offloading affects not only answer quality but also payload size, token cost, and system latency. Proprietary APIs expose little control over model internals or serving behavior, leaving client-side preprocessing as the main practical optimization space for downstream developers. Many such techniques have been proposed for visual offloading, yet their cost-quality impact on commercial cloud VLMs has never been studied. To fill this gap, we present VQABench, the first systematic benchmark that treats client-side input preprocessing as a controlled variable for cloud-VLM-based VQA. We evaluate 12 preprocessing techniques across three VQA datasets and four commercial VLMs from three providers, totaling 95,168 API calls. Our results show that preprocessing is not universally beneficial: its effectiveness depends on the target model, API paradigm, provider token-accounting rule, and task formulation. A poorly selected preprocessing strategy can increase deployment cost or latency while degrading answer accuracy. Overall, our benchmark clarifies when preprocessing helps, when it fails, and why, providing insights to guide future research and real-world deployment of VQA systems.

Enhanced Byzantine-Robust Federated Learning Via Truncated-Quadratic Loss for Heterogeneous Data

Federated learning distributes data among $n$ clients, making it vulnerable to malicious attacks and data heterogeneity, which together pose challenges for robust learning. To tackle this issue, centered clipping and Huber aggregators have been exploited for Byzantine robustness. In this paper, we first demonstrate their equivalence via convex conjugate theory, and show that they can yield biased solutions in the presence of outliers, leading to failure under high data heterogeneity and a substantial fraction of outliers. Next, we propose a new robust aggregation rule that utilizes the truncated-quadratic (TQ) loss, effectively mitigating the biases of existing methods, such as centered clipping and Huber aggregators. We show that our aggregator achieves order-optimal Byzantine-robust learning under nonconvex loss functions and heterogeneous data, ultimately enhancing the reliability of federated learning systems. Additionally, we provide a robust deviation estimation strategy for TQ, demonstrating its effectiveness. Furthermore, we show that TQ maintains robustness even when only an estimate of the number of Byzantine clients is available. Finally, experimental results on MNIST, Fashion-MNIST, and CIFAR-10, indicate that our aggregator provides better robustness performance than the competing techniques.

SwiftChannel: Algorithm-Hardware Co-Design for Deep Learning-Based 5G Channel Estimation

Channel estimation is crucial in 5G communication networks for optimizing transmission parameters and ensuring reliable, high-speed communication. However, the use of multiple-input and multiple-output (MIMO) and millimeter-wave (mmWave) in 5G networks presents challenges in achieving accurate estimation under strict latency requirements on resource-limited hardware platforms. To address these challenges, we propose SwiftChannel, an algorithm-hardware co-design framework that integrates a hardware-friendly deep learning-based channel estimator with a dedicated accelerator. Our approach employs a convolutional neural network enhanced with a parameter-free attention mechanism, which effectively reconstructs full-resolution spatial-frequency domain channel matrices from low-resolution least squares (LS) estimates. We further develop a multi-stage model compression pipeline combining knowledge distillation, convolution re-parameterization, and quantization-aware training, resulting in substantial model size reduction with negligible accuracy loss. The hardware accelerator, implementing the compressed model and the LS estimator on FPGA platforms using High-level Synthesis (HLS), features a fine-grained pipeline architecture and optimized dataflow strategies. Tested on a Zynq UltraScale+ RFSoC, the accelerator achieves sub-millisecond latency, providing up to 24x speed-up and over 33x improvement in energy efficiency compared to GPU-based solutions. Extensive evaluations demonstrate that the proposed design generalizes not only across various noise levels and user mobilities, but also to a variety of unseen channel profiles, outperforming state-of-the-art baselines. By unifying algorithmic innovation with hardware-aware design, our work presents a future-proof channel estimation solution for 5G MIMO systems.

ViM-Q: Scalable Algorithm-Hardware Co-Design for Vision Mamba Model Inference on FPGA

Vision Mamba (ViM) models offer a compelling efficiency advantage over Transformers by leveraging the linear complexity of State Space Models (SSMs), yet efficiently deploying them on FPGAs remains challenging. Linear layers struggle with dynamic activation outliers that render static quantization ineffective, while uniform quantization fails to capture the weight distribution at low bit-widths. Furthermore, while associative scan accelerates SSMs on GPUs, its memory access patterns are misaligned with the streaming dataflow required by FPGAs. To address these challenges, we present ViM-Q, a scalable algorithm-hardware co-design for end-to-end ViM inference on the edge. We introduce a hardware-aware quantization scheme combining dynamic per-token activation quantization and per-channel smoothing to mitigate outliers, alongside a custom 4-bit per-block Additive Power-of-Two (APoT) weight quantization. The models are deployed on a runtime-parameterizable FPGA accelerator featuring a linear engine employing a Lookup-Table (LUT) unit to replace multiplications with shift-add operations, and a fine-grained pipelined SSM engine that parallelizes the state dimension while preserving sequential recurrence. Crucially, the hardware supports runtime configuration, adapting to diverse dimensions and input resolutions across the ViM family. Implemented on an AMD ZCU102 FPGA, ViM-Q achieves an average 4.96x speedup and 59.8x energy efficiency gain over a quantized NVIDIA RTX 3090 GPU baseline for low-batch inference on ViM-tiny. This co-design shows a viable path for deploying ViM models on resource-constrained edge devices.

Beyond Models! Explainable Data Valuation and Metric Adaption for Recommendation

User behavior records serve as the foundation for recommender systems. While the behavior data exhibits ease of acquisition, it often suffers from varying quality. Current methods employ data valuation to discern high-quality data from low-quality data. However, they tend to employ black-box design, lacking transparency and interpretability. Besides, they are typically tailored to specific evaluation metrics, leading to limited generality across various tasks. To overcome these issues, we propose an explainable and versatile framework DVR which can enhance the efficiency of data utilization tailored to any requirements of the model architectures and evaluation metrics. For explainable data valuation, a data valuator is presented to evaluate the data quality via calculating its Shapley value from the game-theoretic perspective, ensuring robust mathematical properties and reliability. In order to accommodate various evaluation metrics, including differentiable and non-differentiable ones, a metric adapter is devised based on reinforcement learning, where a metric is treated as the reinforcement reward that guides model optimization. Extensive experiments conducted on various benchmarks verify that our framework can improve the performance of current recommendation algorithms on various metrics including ranking accuracy, diversity, and fairness. Specifically, our framework achieves up to 34.7\% improvements over existing methods in terms of representative NDCG metric. The code is available at https://github.com/renqii/DVR.

Communication-Efficient and Privacy-Adaptable Mechanism for Federated Learning

Training machine learning models on decentralized private data via federated learning (FL) poses two key challenges: communication efficiency and privacy protection. In this work, we address these challenges within the trusted aggregator model by introducing a novel approach called the Communication-Efficient and Privacy-Adaptable Mechanism (CEPAM), achieving both objectives simultaneously. In particular, CEPAM leverages the rejection-sampled universal quantizer (RSUQ), a construction of randomized vector quantizer whose resulting distortion is equivalent to a prescribed noise, such as Gaussian or Laplace noise, enabling joint differential privacy and compression. Moreover, we analyze the trade-offs among user privacy, global utility, and transmission rate of CEPAM by defining appropriate metrics for FL with differential privacy and compression. Our CEPAM provides the additional benefit of privacy adaptability, allowing clients and the server to customize privacy protection based on required accuracy and protection. We assess CEPAM's utility performance using MNIST dataset, demonstrating that CEPAM surpasses baseline models in terms of learning accuracy.

EmbedGenius: Towards Automated Software Development for Generic Embedded IoT Systems

Embedded IoT system development is crucial for enabling seamless connectivity and functionality across a wide range of applications. However, such a complex process requires cross-domain knowledge of hardware and software and hence often necessitates direct developer involvement, making it labor-intensive, time-consuming, and error-prone. To address this challenge, this paper introduces EmbedGenius, the first fully automated software development platform for general-purpose embedded IoT systems. The key idea is to leverage the reasoning ability of Large Language Models (LLMs) and embedded system expertise to automate the hardware-in-the-loop development process. The main methods include a component-aware library resolution method for addressing hardware dependencies, a library knowledge generation method that injects utility domain knowledge into LLMs, and an auto-programming method that ensures successful deployment. We evaluate EmbedGenius's performance across 71 modules and four mainstream embedded development platforms with over 350 IoT tasks. Experimental results show that EmbedGenius can generate codes with an accuracy of 95.7% and complete tasks with a success rate of 86.5%, surpassing human-in-the-loop baselines by 15.6%--37.7% and 25.5%--53.4%, respectively. We also show EmbedGenius's potential through case studies in environmental monitoring and remote control systems development.

TransCompressor: LLM-Powered Multimodal Data Compression for Smart Transportation

The incorporation of Large Language Models (LLMs) into smart transportation systems has paved the way for improving data management and operational efficiency. This study introduces TransCompressor, a novel framework that leverages LLMs for efficient compression and decompression of multimodal transportation sensor data. TransCompressor has undergone thorough evaluation with diverse sensor data types, including barometer, speed, and altitude measurements, across various transportation modes like buses, taxis, and MTRs. Comprehensive evaluation illustrates the effectiveness of TransCompressor in reconstructing transportation sensor data at different compression ratios. The results highlight that, with well-crafted prompts, LLMs can utilize their vast knowledge base to contribute to data compression processes, enhancing data storage, analysis, and retrieval in smart transportation settings.

EarDA: Towards Accurate and Data-Efficient Earable Activity Sensing

In the realm of smart sensing with the Internet of Things, earable devices are empowered with the capability of multi-modality sensing and intelligence of context-aware computing, leading to its wide usage in Human Activity Recognition (HAR). Nonetheless, unlike the movements captured by Inertial Measurement Unit (IMU) sensors placed on the upper or lower body, those motion signals obtained from earable devices show significant changes in amplitudes and patterns, especially in the presence of dynamic and unpredictable head movements, posing a significant challenge for activity classification. In this work, we present EarDA, an adversarial-based domain adaptation system to extract the domain-independent features across different sensor locations. Moreover, while most deep learning methods commonly rely on training with substantial amounts of labeled data to offer good accuracy, the proposed scheme can release the potential usage of publicly available smartphone-based IMU datasets. Furthermore, we explore the feasibility of applying a filter-based data processing method to mitigate the impact of head movement. EarDA, the proposed system, enables more data-efficient and accurate activity sensing. It achieves an accuracy of 88.8% under HAR task, demonstrating a significant 43% improvement over methods without domain adaptation. This clearly showcases its effectiveness in mitigating domain gaps.

Are You Being Tracked? Discover the Power of Zero-Shot Trajectory Tracing with LLMs!

There is a burgeoning discussion around the capabilities of Large Language Models (LLMs) in acting as fundamental components that can be seamlessly incorporated into Artificial Intelligence of Things (AIoT) to interpret complex trajectories. This study introduces LLMTrack, a model that illustrates how LLMs can be leveraged for Zero-Shot Trajectory Recognition by employing a novel single-prompt technique that combines role-play and think step-by-step methodologies with unprocessed Inertial Measurement Unit (IMU) data. We evaluate the model using real-world datasets designed to challenge it with distinct trajectories characterized by indoor and outdoor scenarios. In both test scenarios, LLMTrack not only meets but exceeds the performance benchmarks set by traditional machine learning approaches and even contemporary state-of-the-art deep learning models, all without the requirement of training on specialized datasets. The results of our research suggest that, with strategically designed prompts, LLMs can tap into their extensive knowledge base and are well-equipped to analyze raw sensor data with remarkable effectiveness.