Compared to traditional Vision Transformers (ViT), Mixture-of-Experts ViTs (MoE-ViTs) are introduced to scale model size without a proportional increase in computational complexity, making them a new research focus. Given the high performance and reconfigurability, field-programmable gate array (FPGA)-based accelerators for MoE-ViTs emerge, delivering substantial gains over general-purpose processors. However, existing accelerators often fall short of efficiently managing the highly dynamic and sparse computation patterns, resulting in suboptimal tradeoffs between resource utilization and performance. To address the inefficiencies in deploying MoE-ViTs on FPGAs, we present MoE-Sched, a novel end-to-end accelerator that embraces a scheduling-centric design philosophy. Rather than optimizing isolated kernels, MoE-Sched coordinates multilevel scheduling, from fine-grained intrakernel streaming to module reuse and multidie mapping, to holistically balance latency, bandwidth (BW), and resource usage. We further integrate a hardware-aware quantization scheme tailored for streaming attention and sparse expert execution, preserving accuracy while minimizing overhead. Experimental results demonstrate that our accelerator achieves nearly 100 frames/s on M3ViT-tiny, a$3.13\times $improvement in throughput, and over 75% energy reduction compared to state-of-the-art (SOTA) FPGA MoE accelerators, while maintaining less than 1% accuracy loss across vision benchmarks. Our implementation will be open-sourced.

Compared to traditional Vision Transformers (ViT), Mixture-of-Experts Vision Transformers (MoE-ViT) are introduced to scale model size without a proportional increase in computational complexity, making them a new research focus. Given the high performance and reconfigurability, FPGA-based accelerators for MoE-ViT emerge, delivering substantial gains over general-purpose processors. However, existing accelerators often fall short of fully exploring the design space, leading to suboptimal trade-offs between resource utilization and performance. To overcome this problem, we introduce UbiMoE, a novel end-to-end FPGA accelerator tailored for MoE-ViT. Leveraging the unique computational and memory access patterns of MoE-ViTs, we develop a latency-optimized streaming attention kernel and a resource-efficient reusable linear kernel, effectively balancing performance and resource consumption. To further enhance design efficiency, we propose a two-stage heuristic search algorithm that optimally tunes hardware parameters for various FPGA resource constraints. Compared to state-of-the-art (SOTA) FPGA designs, UbiMoE achieves 1.34× and 3.35× throughput improvements for MoE-ViT on Xilinx ZCU102 and Alveo U280 platforms, respectively, while enhancing energy efficiency by 1.75× and 1.54×. Our implementation is available at https://github.com/DJ000011/UbiMoE.

This Paper introduces a novel Enhanced Visual Retrieval and Analysis for Charging Efficiency (EVRACE) system, which presents a novel two-stage framework for visual retrieval and advanced analytics for massive charging curves of electric vehicles. The EVRACE is becoming an important system for enhancing battery performance management, detecting and mitigating charging anomalies, and informing strategic decision-making for stakeholders in the EV ecosystem. EVRACE utilizes deep time-series clustering to categorize charging curves with distinct characteristics and retrieves the most relevant cluster and the top k curves for comprehensive data analysis and risk assessment. This innovative framework substantially enhances retrieval efficiency, mitigates computational complexity, and offers a robust solution for real-time processing of large-scale charging data. Validation using real-world datasets demonstrates that the EVRACE system significantly improves retrieval efficiency and achieves high accuracy in anomaly detection compared to traditional methods.

Learning from demonstration (LfD) is a promising method for robots to learn and generalize human-like skills. It has the advantages of high programming efficiency, easy optimization, and non-professionals can also operate. There is a lot of research work that learn motion trajectories and stiffness curves from human demonstrations simutaneously to make the robot compliant, but previous work rarely consider the changes of environment. In this article, we propose an adaptive stiffness method that enables the robot to learn motion and stiffness trajectories from a single demonstration. When the environment changes, it can spontaneously tune the stiffness according to environmental feedback to ensure the smoothness of the task. Thus the robot has the ability to adapt to environmental changes. We first proved the theoretical feasibility of the method, and then we conducted physical experiments on the Baxter robot to verify the effectiveness of the proposed method.