Related papers: POS-ISP: Pipeline Optimization at the Sequence Lev…

Multi-task model training has been adopted to enable a single deep neural network model (often a large language model) to handle multiple tasks (e.g., question answering and text summarization). Multi-task training commonly receives input…

Pipeline parallelism (PP) when training neural networks enables larger models to be partitioned spatially, leading to both lower network communication and overall higher hardware utilization. Unfortunately, to preserve the statistical…

Autoscaling GPU inference workloads in Kubernetes remains challenging due to the reactive and threshold-based nature of default mechanisms such as the Horizontal Pod Autoscaler (HPA), which struggle under dynamic and bursty traffic patterns…

High-resolution images are prevalent in various applications, such as autonomous driving and computer-aided diagnosis. However, training neural networks on such images is computationally challenging and easily leads to out-of-memory errors…

Intelligent Reflecting Surface (IRS) is a promising technology for next generation wireless networks. Despite substantial research in IRS-aided communications, the assumed antenna and channel models are typically simplified without…

This paper introduces scalable, sampling-based algorithms that optimize trained neural networks with ReLU activations. We first propose an iterative algorithm that takes advantage of the piecewise linear structure of ReLU neural networks…

This paper studies spatial capacity in a stochastic wireless ad hoc network, where multi-stage probing and data transmission are sequentially performed. We propose a novel signal-to-interference-ratio (SIR) threshold based scheduling…

The physical layer security (PLS) is investigated for reconfigurable intelligent surface (RIS) assisted wireless networks, where a source transmits its confidential information to a legitimate destination with the aid of a single small RIS…

Feature preprocessing, which transforms raw input features into numerical representations, is a crucial step in automated machine learning (AutoML) systems. However, the existing systems often have a very small search space for feature…

Region proposal is critical for object detection while it usually poses a bottleneck in improving the computation efficiency on traditional control-flow architectures. We have observed region proposal tasks are potentially suitable for…

Recent considerations for reconfigurable intelligent surfaces (RISs) assume that RISs can convey information by reflection without the need of transmit radio frequency chains, which, however, is a challenging task. In this paper, we propose…

Large reasoning models improve with more test-time computation, but often overthink, producing unnecessarily long chains-of-thought that raise cost without improving accuracy. Prior reinforcement learning approaches typically rely on a…

End-to-end autonomous driving is typically built upon imitation learning (IL), yet its performance is constrained by the quality of human demonstrations. To overcome this limitation, recent methods incorporate reinforcement learning (RL)…

Intelligent reflecting surface (IRS) is an emerging technology to enhance the spectral and energy efficiency of wireless communications cost-effectively. This letter considers a new multi-IRS aided wireless network where a cascaded…

It is a challenging task to train large DNN models on sophisticated GPU platforms with diversified interconnect capabilities. Recently, pipelined training has been proposed as an effective approach for improving device utilization. However,…

Vision-Language Models require efficient adaptation to continually emerging downstream tasks. While Parameter-Efficient Fine-Tuning mitigates catastrophic forgetting, assigning isolated modules per task leads to parameter explosion.…

Integrated sensing, computation, and communication (ISCC) has been recently considered as a promising technique for beyond 5G systems. In ISCC systems, the competition for communication and computation resources between sensing tasks for…

We introduce BriskStream, an in-memory data stream processing system (DSPSs) specifically designed for modern shared-memory multicore architectures. BriskStream's key contribution is an execution plan optimization paradigm, namely RLAS,…

Spiking Neural Networks (SNNs) have gained popularity due to their high energy efficiency. Prior works have proposed various methods for training SNNs, including backpropagation-based methods. Training SNNs is computationally expensive…

We optimize pipeline parallelism for deep neural network (DNN) inference by partitioning model graphs into $k$ stages and minimizing the running time of the bottleneck stage, including communication. We give practical and effective…