arXiv subjects

Xinglu Wang

Publications and source records attributed to Xinglu Wang.

10 recordsLinked to original sources

arXiv2026

ReviveMoE: Fast Recovery for Hardware Failures in Large-Scale MoE LLM Inference Deployments

As LLM deployments scale over more hardware, the probability of a single failure in a system increases significantly, and cloud operators must consider robust countermeasures to handle these inevitable failures. A common recovery approach is to simply restart the LLM serving instance; however, this is costly in model-as-a-service (MaaS) inference settings, where reloading model weights and recompiling computation graphs can introduce significant delays to incoming requests. We propose ReviveMoE, a method for rapid failure recovery in large-scale LLM deployments without restarting the serving instance. ReviveMoE is designed to support both the traditional LLM architecture, which collocates MoE and attention on the same hardware, and the disaggregated architectures, which separate MoE from attention. Integrated into Huawei Cloud's MaaS, ReviveMoE is built on top of Huawei's xDeepServe serving platform and the XCCL communications library.

cs.DC↗

arXiv2025

HFX: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling

Large language model (LLM) serving faces the dual challenge of meeting strict user-specific service-level objectives (SLOs) while minimizing computational cost under dynamic, multi-task workloads. Existing approaches either rely on static scheduling policies or focus on single-task settings, limiting their applicability in real-world deployments with heterogeneous requests, variable prompt lengths, and elastic scaling requirements. We present HFX, a production LLM serving system that jointly optimizes request scheduling and elastic scaling across model replicas to satisfy diverse SLOs. HFX introduces a \textbf{scheduler} that performs proactive budget estimation and prioritization to ensure SLO compliance for both new and in-flight requests. HFX also integrates a \textbf{scaler} that supports fast device-to-device (D2D) weight transfer, reducing cold-start latency. Additionally, the system supports both colocated and disaggregated prefill/decode deployments, enabling adaptation to diverse workload patterns and cloud environments. Through extensive experiments on multi-task workloads, we demonstrate consistently higher SLO attainment, lower end-to-end latency, and lower NPU usage cost by up to 4.44$\times$, 65.82\%, and 49.81\%, respectively, compared to state-of-the-art systems. Our results highlight the effectiveness of SLO-aware scheduling and scaling in practical LLM serving, providing a robust framework for cost-efficient and SLO-compliant deployments.

cs.DC↗

arXiv2024

Efficiently Serving Large Multimodal Models Using EPD Disaggregation

Large Multimodal Models (LMMs) extend Large Language Models (LLMs) by handling diverse inputs such as images, audio, and video, but at the cost of adding a multimodal encoding stage that increases both computational and memory overhead. This step negatively affects key Service Level Objectives (SLOs), such as time to first token (TTFT) and time per output token (TPOT). We introduce Encode-Prefill-Decode (EPD) Disaggregation, a novel framework that separates the encoding, prefill, and decode stages onto dedicated resources. Unlike current systems, which bundle encoding and prefill together, our approach decouples these steps, unlocking new opportunities and optimizations. These include a mechanism to cache multimedia tokens for efficient transfer, a novel way to parallelize the encoding load within a request, a module for optimal resource allocation for disaggregated serving, and a novel role-switching method to handle changing workload characteristics. Experimental evaluations with popular LMMs show substantial gains in memory efficiency (up to 15x lower peak memory utilization), batch sizes (up to 22x larger), 10x more images per request, and 2.2x larger KV caches. Furthermore, it leads to significant improvements in SLO attainment (up to 90-100% improvement) and TTFT (up to 71% reduction), compared to systems that do not disaggregate. The code is available at https://github.com/vbdi/epdserve.

cs.DC↗

arXiv2024

Artificial Intelligence for Operations Research: Revolutionizing the Operations Research Process

The rapid advancement of artificial intelligence (AI) techniques has opened up new opportunities to revolutionize various fields, including operations research (OR). This survey paper explores the integration of AI within the OR process (AI4OR) to enhance its effectiveness and efficiency across multiple stages, such as parameter generation, model formulation, and model optimization. By providing a comprehensive overview of the state-of-the-art and examining the potential of AI to transform OR, this paper aims to inspire further research and innovation in the development of AI-enhanced OR methods and tools. The synergy between AI and OR is poised to drive significant advancements and novel solutions in a multitude of domains, ultimately leading to more effective and efficient decision-making.

math.OC↗

arXiv2023

ETran: Energy-Based Transferability Estimation

This paper addresses the problem of ranking pre-trained models for object detection and image classification. Selecting the best pre-trained model by fine-tuning is an expensive and time-consuming task. Previous works have proposed transferability estimation based on features extracted by the pre-trained models. We argue that quantifying whether the target dataset is in-distribution (IND) or out-of-distribution (OOD) for the pre-trained model is an important factor in the transferability estimation. To this end, we propose ETran, an energy-based transferability assessment metric, which includes three scores: 1) energy score, 2) classification score, and 3) regression score. We use energy-based models to determine whether the target dataset is OOD or IND for the pre-trained model. In contrast to the prior works, ETran is applicable to a wide range of tasks including classification, regression, and object detection (classification+regression). This is the first work that proposes transferability estimation for object detection task. Our extensive experiments on four benchmarks and two tasks show that ETran outperforms previous works on object detection and classification benchmarks by an average of 21% and 12%, respectively, and achieves SOTA in transferability assessment.

cs.CV↗

arXiv2022

Objects Matter: Learning Object Relation Graph for Robust Camera Relocalization

Visual relocalization aims to estimate the pose of a camera from one or more images. In recent years deep learning based pose regression methods have attracted many attentions. They feature predicting the absolute poses without relying on any prior built maps or stored images, making the relocalization very efficient. However, robust relocalization under environments with complex appearance changes and real dynamics remains very challenging. In this paper, we propose to enhance the distinctiveness of the image features by extracting the deep relationship among objects. In particular, we extract objects in the image and construct a deep object relation graph (ORG) to incorporate the semantic connections and relative spatial clues of the objects. We integrate our ORG module into several popular pose regression models. Extensive experiments on various public indoor and outdoor datasets demonstrate that our method improves the performance significantly and outperforms the previous approaches.

cs.CV↗

arXiv2022

Large-Area Intercalated 2D-Pb/Graphene Heterostructure as a Platform for Generating Spin-Orbit Torque

A scalable platform to synthesize ultrathin heavy metals may enable high efficiency charge-to-spin conversion for next-generation spintronics. Here we report the synthesis of air-stable, epitaxially registered monolayer Pb underneath bilayer graphene on SiC (0001) by confinement heteroepitaxy (CHet). Diffraction, spectroscopy, and microscopy reveal CHet-based Pb intercalation predominantly exhibits a mottled hexagonal superstructure due to an ordered network of Frenkel-Kontorova-like domain walls. The system's air stability enables ex-situ spin torque ferromagnetic resonance (ST-FMR) measurements that demonstrate charge-to-spin conversion in graphene/Pb/ferromagnet heterostructures with a 1.5x increase in the effective field ratio compared to control samples.

cond-mat.mtrl-sci↗

arXiv2022

Fair and efficient contribution valuation for vertical federated learning

Federated learning is an emerging technology for training machine learning models across decentralized data sources without sharing data. Vertical federated learning, also known as feature-based federated learning, applies to scenarios where data sources have the same sample IDs but different feature sets. To ensure fairness among data owners, it is critical to objectively assess the contributions from different data sources and compensate the corresponding data owners accordingly. The Shapley value is a provably fair contribution valuation metric originating from cooperative game theory. However, its straight-forward computation requires extensively retraining a model on each potential combination of data sources, leading to prohibitively high communication and computation overheads due to multiple rounds of federated learning. To tackle this challenge, we propose a contribution valuation metric called vertical federated Shapley value (VerFedSV) based on the classic Shapley value. We show that VerFedSV not only satisfies many desirable properties of fairness but is also efficient to compute. Moreover, VerFedSV can be adapted to both synchronous and asynchronous vertical federated learning algorithms. Both theoretical analysis and extensive experimental results demonstrate the fairness, efficiency, adaptability, and effectiveness of VerFedSV.

cs.LG↗

arXiv2020

Person Re-identification with Adversarial Triplet Embedding

Person re-identification is an important task and has widespread applications in video surveillance for public security. In the past few years, deep learning network with triplet loss has become popular for this problem. However, the triplet loss usually suffers from poor local optimal and relies heavily on the strategy of hard example mining. In this paper, we propose to address this problem with a new deep metric learning method called Adversarial Triplet Embedding (ATE), in which we simultaneously generate adversarial triplets and discriminative feature embedding in an unified framework. In particular, adversarial triplets are generated by introducing adversarial perturbations into the training process. This adversarial game is converted into a minimax problem so as to have an optimal solution from the theoretical view. Extensive experiments on several benchmark datasets demonstrate the effectiveness of the approach against the state-of-the-art literature.

cs.CV↗

arXiv2020

Adversarial Multi-scale Feature Learning for Person Re-identification

Person Re-identification (Person ReID) is an important topic in intelligent surveillance and computer vision. It aims to accurately measure visual similarities between person images for determining whether two images correspond to the same person. The key to accurately measure visual similarities is learning discriminative features, which not only captures clues from different spatial scales, but also jointly inferences on multiple scales, with the ability to determine reliability and ID-relativity of each clue. To achieve these goals, we propose to improve Person ReID system performance from two perspective: \textbf{1).} Multi-scale feature learning (MSFL), which consists of Cross-scale information propagation (CSIP) and Multi-scale feature fusion (MSFF), to dynamically fuse features cross different scales.\textbf{2).} Multi-scale gradient regularizor (MSGR), to emphasize ID-related factors and ignore irrelevant factors in an adversarial manner. Combining MSFL and MSGR, our method achieves the state-of-the-art performance on four commonly used person-ReID datasets with neglectable test-time computation overhead.

cs.CV↗