Related papers: Improving PIE's performance over high-delay paths

Mixture-of-Experts (MoE) has successfully scaled up models while maintaining nearly constant computing costs. By employing a gating network to route input tokens, it selectively activates a subset of expert networks to process the…

Active queue control aims to improve the overall communication network throughput while providing lower delay and small packet loss rate. The basic idea is to actively trigger packet dropping (or marking provided by explicit congestion…

Explicit Congestion Notification (ECN)-based congestion control schemes have been widely adopted in high-speed data center networks (DCNs), where the ECN marking threshold plays a determinant role in guaranteeing a packet lossless DCN.…

Operating cloud service infrastructures requires high energy efficiency while ensuring a satisfactory service level. Motivated by data centers, we consider a workload routing and server speed control policy applicable to the system…

Mixture-of-Experts (MoE) architectures offer the promise of larger model capacity without the prohibitive costs of fully dense designs. However, in real-world inference serving, load skew across experts often leads to suboptimal device…

Mixture-of-Experts (MoE) architectures scale language models by activating only a subset of specialized expert networks for each input token, thereby reducing the number of floating-point operations. However, the growing size of modern MoE…

This dissertation is a study on the design and analysis of novel, optimal routing and rate control algorithms in wireless, mobile communication networks. Congestion control and routing algorithms upto now have been designed and optimized…

Mixture-of-Experts (MoE) models are designed to enhance the efficiency of large language models (LLMs) without proportionally increasing the computational demands. However, their deployment on edge devices still faces significant challenges…

Immersive media streaming, especially virtual reality (VR)/360-degree video streaming which is very bandwidth demanding, has become more and more popular due to the rapid growth of the multimedia and networking deployments. To better…

The mixture of experts (MoE) model is a sparse variant of large language models (LLMs), designed to hold a better balance between intelligent capability and computational overhead. Despite its benefits, MoE is still too expensive to deploy…

Computing-In-Memory (CIM) offers a potential solution to the memory wall issue and can achieve high energy efficiency by minimizing data movement, making it a promising architecture for edge AI devices. Lightweight models like MobileNet and…

In this work, the achievable rate of three-node relay systems with selection relaying under statistical delay constraints, imposed on the limitations of the maximum end-to-end delay violation probabilities, is investigated. It is assumed…

The recent detection transformer (DETR) has advanced object detection, but its application on resource-constrained devices requires massive computation and memory resources. Quantization stands out as a solution by representing the network…

The proliferation of large language models (LLMs) has driven the adoption of Mixture-of-Experts (MoE) architectures as a promising solution to scale model capacity while controlling computational costs. However, deploying MoE models in…

As computation shifts from the cloud to the edge to reduce processing latency and network traffic, the resulting Computing Continuum (CC) creates a dynamic environment where meeting strict Quality of Service (QoS) requirements and avoiding…

This work proposes adaptive buffer-aided distributed space-time coding schemes and algorithms with feedback for wireless networks equipped with buffer-aided relays. The proposed schemes employ a maximum likelihood receiver at the…

Mixture-of-Experts models have become a dominant architecture for scaling Large Language Models by activating only a sparse subset of experts per token. However, latency-critical MoE inference faces a fundamental tension: while expert…

Large language model (LLM) serving is becoming an increasingly critical workload for cloud providers. Existing LLM serving systems focus on interactive requests, such as chatbots and coding assistants, with tight latency SLO requirements.…

We consider the problem of routing packets across a multi-hop network consisting of multiple sources of traffic and wireless links while ensuring bounded expected delay. Each packet transmission can be overheard by a random subset of…

We develop link rate control policies to minimize the queueing delay of packets in overloaded networks. We show that increasing link rates does not guarantee delay reduction during overload. We consider a fluid queueing model that…