Related papers: SA-Kura: An Energy-Efficient Systolic Array Accele…

Multi-pod systolic arrays are emerging as the architecture of choice in DNN inference accelerators. Despite their potential, designing multi-pod systolic arrays to maximize effective throughput/Watt (i.e., throughput/Watt adjusted when…

Time Series Analysis (TSA) is a critical workload to extract valuable information from collections of sequential data, e.g., detecting anomalies in electrocardiograms. Subsequence Dynamic Time Warping (sDTW) is the state-of-the-art…

With the growing performance requirements on networked applications, there is a new trend of offloading stateful network applications to SmartNICs to improve performance and reduce the total cost of ownership. However, offloading stateful…

Spiking Neural Networks (SNNs) have gained significant attention in edge computing due to their low power consumption and computational efficiency. However, existing implementations either use conventional System on Chip (SoC) architectures…

Integration of intermittent renewable energy sources in modern power systems is increasing very fast. Replacement of synchronous generators with zero-to-low variable renewables substantially decreases the system inertia. In a large system,…

Edge inference for large language models (LLM) offers secure, low-latency, and cost-effective inference solutions. We emphasize that an edge accelerator should achieve high area efficiency and minimize external memory access (EMA) during…

Datacenter servers are increasingly heterogeneous: from x86 host CPUs, to ARM or RISC-V CPUs in NICs/SSDs, to FPGAs. Previous works have demonstrated that migrating application execution at run-time across heterogeneous-ISA CPUs can yield…

Seismic data interpolation is a critical pre-processing step for improving seismic imaging quality and remains a focus of academic innovation. To address the computational inefficiencies caused by extensive iterative resampling in current…

Neural network accelerators have been widely applied to edge devices for complex tasks like object tracking, image recognition, etc. Previous works have explored the quantization technologies in related lightweight accelerator designs to…

This paper introduces the sparse periodic systolic (SPS) dataflow, which advances the state-of-the-art hardware accelerator for supporting lightweight neural networks. Specifically, the SPS dataflow enables a novel hardware design approach…

Attention mechanisms, which enable a neural network to accurately focus on all the relevant elements of the input, have become an essential component to improve the performance of deep neural networks. There are mainly two attention…

Deep learning-driven superresolution (SR) outperforms traditional techniques but also faces the challenge of high complexity and memory bandwidth. This challenge leads many accelerators to opt for simpler and shallow models like FSRCNN,…

While diffusion models have achieved great success in the field of video generation, this progress is accompanied by a rapidly escalating computational burden. Among the existing acceleration methods, Feature Caching is popular due to its…

We propose an efficient deep learning method for single image defocus deblurring (SIDD) by further exploring inverse kernel properties. Although the current inverse kernel method, i.e., kernel-sharing parallel atrous convolution (KPAC), can…

Existing attention accelerators often trade exact softmax semantics, depend on fused Tensor Core kernels, or incur sequential depth that limits FP32 throughput on long sequences. We present \textbf{ELSA}, an algorithmic reformulation of…

Deep learning has achieved some success in addressing the challenge of cloud removal in optical satellite images, by fusing with synthetic aperture radar (SAR) images. Recently, diffusion models have emerged as powerful tools for cloud…

Improving the efficiency of edge detection in embedded applications, such as UAV control, is critical for reducing system cost and power dissipation. Field programmable gate arrays (FPGA) are a good platform for making improvements because…

The increasing complexity of deep neural networks (DNNs) poses significant challenges for edge inference deployment due to resource and power constraints of edge devices. Recent works on unary-based matrix multiplication hardware aim to…

We investigate synchronization in networks of Kuramoto oscillators with inertia. More specifically, we introduce a rewiring algorithm consisting basically in a {\em hill climb} scheme in which the edges of the network are swapped in order…

Recent advancements in diffusion-based generative priors have enabled visually plausible image compression at extremely low bit rates. However, existing approaches suffer from slow sampling processes and suboptimal bit allocation due to…