arXiv subjects

Quan Chen

Publications and source records attributed to Quan Chen.

At least 73 records · Page 4Linked to original sources

arXiv2025

CoMoE: Collaborative Optimization of Expert Aggregation and Offloading for MoE-based LLMs at Edge

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 resource-constrained mobile edge computing environments presents significant challenges due to their large memory footprint and dynamic expert activation patterns. To address these challenges, we propose a novel dynamic resource-aware collaborative optimization framework that jointly optimizes expert aggregation granularity and offloading strategies based on real-time device resource states, network conditions, and input characteristics in mobile edge environments, denoted as CoMoE. In CoMoE, we first systematically analyze existing expert aggregation techniques, including expert parameter merging,knowledge distillation,and parameter sharing decomposition, identifying their limitations in dynamic mobile environments.We then investigate expert offloading strategies encompassing expert prediction and prefetching, expert caching and scheduling, and multi-tier storage architectures, revealing the interdependencies between routing decisions and offloading performance.The CoMoE incorporates adaptive scheduling mechanisms that respond to user mobility and varying network conditions, enabling efficient MoE deployment across heterogeneous edge devices. Extensive experiments on real mobile edge testbeds demonstrate that CoMoE achieves approximately 70% reduction in memory usage compared to baseline methods, 10.5% lower inference latency than existing expert offloading techniques, while maintaining model performance stability. For large-scale MoE models (e.g,7.4B-parameter Switch-Base-128), the CoMoE reduces memory requirements from 15.6GB to 4.7GB, enabling deployment on resource-constrained mobile edge devices that previously could only support much smaller models.

cs.NI↗

arXiv2025

Kairos: Low-latency Multi-Agent Serving with Shared LLMs and Excessive Loads in the Public Cloud

Multi-agent applications utilize the advanced capabilities of large language models (LLMs) for intricate task completion through agent collaboration in a workflow. Under this situation, requests from different agents usually access the same shared LLM to perform different kinds of tasks, forcing the shared LLM to suffer excessive loads. However, existing works have low serving performance for these multi-agent applications, mainly due to the ignorance of inter-agent latency and resource differences for request scheduling. We therefore propose Kairos, a multi-agent orchestration system that optimizes end-to-end latency for multi-agent applications. Kairos consists of a workflow orchestrator, a workflow-aware priority scheduler, and a memory-aware dispatcher. The orchestrator collects agent-specific information for online workflow analysis. The scheduler decides the serving priority of the requests based on their latency characteristics to reduce the overall queuing. The dispatcher dispatches the requests to different LLM instances based on their memory demands to avoid GPU overloading. Experimental results show that Kairos reduces end-to-end latency by 17.8% to 28.4% compared to state-of-the-art works.

cs.DC↗

arXiv2025

Music Grounding by Short Video

Adding proper background music helps complete a short video to be shared. Previous work tackles the task by video-to-music retrieval (V2MR), aiming to find the most suitable music track from a collection to match the content of a given query video. In practice, however, music tracks are typically much longer than the query video, necessitating (manual) trimming of the retrieved music to a shorter segment that matches the video duration. In order to bridge the gap between the practical need for music moment localization and V2MR, we propose a new task termed Music Grounding by Short Video (MGSV). To tackle the new task, we introduce a new benchmark, MGSV-EC, which comprises a diverse set of 53k short videos associated with 35k different music moments from 4k unique music tracks. Furthermore, we develop a new baseline method, MaDe, which performs both video-to-music matching and music moment detection within a unified end-to-end deep network. Extensive experiments on MGSV-EC not only highlight the challenging nature of MGSV but also set MaDe as a strong baseline.

cs.MM↗

arXiv2025

ASR-enhanced Multimodal Representation Learning for Cross-Domain Product Retrieval

E-commerce is increasingly multimedia-enriched, with products exhibited in a broad-domain manner as images, short videos, or live stream promotions. A unified and vectorized cross-domain production representation is essential. Due to large intra-product variance and high inter-product similarity in the broad-domain scenario, a visual-only representation is inadequate. While Automatic Speech Recognition (ASR) text derived from the short or live-stream videos is readily accessible, how to de-noise the excessively noisy text for multimodal representation learning is mostly untouched. We propose ASR-enhanced Multimodal Product Representation Learning (AMPere). In order to extract product-specific information from the raw ASR text, AMPere uses an easy-to-implement LLM-based ASR text summarizer. The LLM-summarized text, together with visual data, is then fed into a multi-branch network to generate compact multimodal embeddings. Extensive experiments on a large-scale tri-domain dataset verify the effectiveness of AMPere in obtaining a unified multimodal product representation that clearly improves cross-domain product retrieval.

cs.MM↗

arXiv2025

Efficient Unified Caching for Accelerating Heterogeneous AI Workloads

Modern AI clusters, which host diverse workloads like data pre-processing, training and inference, often store the large-volume data in cloud storage and employ caching frameworks to facilitate remote data access. To avoid code-intrusion complexity and minimize cache space wastage, it is desirable to maintain a unified cache shared by all the workloads. However, existing cache management strategies, designed for specific workloads, struggle to handle the heterogeneous AI workloads in a cluster -- which usually exhibit heterogeneous access patterns and item storage granularities. In this paper, we propose IGTCache, a unified, high-efficacy cache for modern AI clusters. IGTCache leverages a hierarchical access abstraction, AccessStreamTree, to organize the recent data accesses in a tree structure, facilitating access pattern detection at various granularities. Using this abstraction, IGTCache applies hypothesis testing to categorize data access patterns as sequential, random, or skewed. Based on these detected access patterns and granularities, IGTCache tailors optimal cache management strategies including prefetching, eviction, and space allocation accordingly. Experimental results show that IGTCache increases the cache hit ratio by 55.6% over state-of-the-art caching frameworks, reducing the overall job completion time by 52.2%.

cs.DC↗

arXiv2025

Instruction Learning Paradigms: A Dual Perspective on White-box and Black-box LLMs

Optimizing instructions for large language models (LLMs) is critical for harnessing their full potential in complex and diverse tasks. However, relying solely on white-box approaches demands extensive computational resources and offers limited representational capacity, while black-box models can incur prohibitive financial costs. To address these challenges, we introduce a novel framework that seamlessly merges the strengths of both paradigms. Black-box models provide high-quality, diverse instruction initializations, and white-box models supply fine-grained interpretability through hidden states and output features. By enforcing a semantic similarity constraint, these components fuse into a unified high-dimensional representation that captures deep semantic and structural nuances, enabling an iterative optimization process to refine instruction quality and adaptability. Extensive evaluations across a broad spectrum of tasks-ranging from complex reasoning to cross-lingual generalization-demonstrate that our approach consistently outperforms state-of-the-art baselines. This fusion of black-box initialization with advanced semantic refinement yields a scalable and efficient solution, paving the way for next-generation LLM-driven applications in diverse real-world scenarios. The source code will be released soon.

cs.CL↗

arXiv2025

Advancing LLM Safe Alignment with Safety Representation Ranking

The rapid advancement of large language models (LLMs) has demonstrated milestone success in a variety of tasks, yet their potential for generating harmful content has raised significant safety concerns. Existing safety evaluation approaches typically operate directly on textual responses, overlooking the rich information embedded in the model's internal representations. In this paper, we propose Safety Representation Ranking (SRR), a listwise ranking framework that selects safe responses using hidden states from the LLM itself. SRR encodes both instructions and candidate completions using intermediate transformer representations and ranks candidates via a lightweight similarity-based scorer. Our approach directly leverages internal model states and supervision at the list level to capture subtle safety signals. Experiments across multiple benchmarks show that SRR significantly improves robustness to adversarial prompts. Our code will be available upon publication.

cs.CL↗

arXiv2025

VGNC: Reducing the Overfitting of Sparse-view 3DGS via Validation-guided Gaussian Number Control

Sparse-view 3D reconstruction is a fundamental yet challenging task in practical 3D reconstruction applications. Recently, many methods based on the 3D Gaussian Splatting (3DGS) framework have been proposed to address sparse-view 3D reconstruction. Although these methods have made considerable advancements, they still show significant issues with overfitting. To reduce the overfitting, we introduce VGNC, a novel Validation-guided Gaussian Number Control (VGNC) approach based on generative novel view synthesis (NVS) models. To the best of our knowledge, this is the first attempt to alleviate the overfitting issue of sparse-view 3DGS with generative validation images. Specifically, we first introduce a validation image generation method based on a generative NVS model. We then propose a Gaussian number control strategy that utilizes generated validation images to determine the optimal Gaussian numbers, thereby reducing the issue of overfitting. We conducted detailed experiments on various sparse-view 3DGS baselines and datasets to evaluate the effectiveness of VGNC. Extensive experiments show that our approach not only reduces overfitting but also improves rendering quality on the test set while decreasing the number of Gaussian points. This reduction lowers storage demands and accelerates both training and rendering. The code will be released.

cs.CV↗

arXiv2025

Improving the Serving Performance of Multi-LoRA Large Language Models via Efficient LoRA and KV Cache Management

Multiple Low-Rank Adapters (Multi-LoRAs) are gaining popularity for task-specific Large Language Model (LLM) applications. For multi-LoRA serving, caching hot KV caches and LoRA adapters in high bandwidth memory of accelerations can improve inference performance. However, existing Multi-LoRA inference systems fail to optimize serving performance like Time-To-First-Toke (TTFT), neglecting usage dependencies when caching LoRAs and KVs. We therefore propose FASTLIBRA, a Multi-LoRA caching system to optimize the serving performance. FASTLIBRA comprises a dependency-aware cache manager and a performance-driven cache swapper. The cache manager maintains the usage dependencies between LoRAs and KV caches during the inference with a unified caching pool. The cache swapper determines the swap-in or out of LoRAs and KV caches based on a unified cost model, when the HBM is idle or busy, respectively. Experimental results show that ELORA reduces the TTFT by 63.4% on average, compared to state-of-the-art works.

cs.AR↗

arXiv2025

Orthus: Autoregressive Interleaved Image-Text Generation with Modality-Specific Heads

We introduce Orthus, an autoregressive (AR) transformer that excels in generating images given textual prompts, answering questions based on visual inputs, and even crafting lengthy image-text interleaved contents. Unlike prior arts on unified multimodal modeling, Orthus simultaneously copes with discrete text tokens and continuous image features under the AR modeling principle. The continuous treatment of visual signals minimizes the information loss for both image understanding and generation while the fully AR formulation renders the characterization of the correlation between modalities straightforward. The key mechanism enabling Orthus to leverage these advantages lies in its modality-specific heads -- one regular language modeling (LM) head predicts discrete text tokens and one diffusion head generates continuous image features conditioning on the output of the backbone. We devise an efficient strategy for building Orthus -- by substituting the Vector Quantization (VQ) operation in the existing unified AR model with a soft alternative, introducing a diffusion head, and tuning the added modules to reconstruct images, we can create an Orthus-base model effortlessly (e.g., within mere 72 A100 GPU hours). Orthus-base can further embrace post-training to better model interleaved images and texts. Empirically, Orthus surpasses competing baselines including Show-o and Chameleon across standard benchmarks, achieving a GenEval score of 0.58 and an MME-P score of 1265.8 using 7B parameters. Orthus also shows exceptional mixed-modality generation capabilities, reflecting the potential for handling intricate practical generation tasks.

cs.CV↗

arXiv2025

Scale-adaptive UAV Geo-localization via Height-aware Partition Learning

UAV Geo-Localization faces significant challenges due to the drastic appearance discrepancy between dronecaptured images and satellite views. Existing methods typically assume a consistent scaling factor across views and rely on predefined partition alignment to extract viewpoint-invariant representations through part-level feature construction. However, this scaling assumption often fails in real-world scenarios, where variations in drone flight states lead to scale mismatches between cross-view images, resulting in severe performance degradation. To address this issue, we propose a scale-adaptive partition learning framework that leverages known drone flight height to predict scale factors and dynamically adjust feature extraction. Our key contribution is a height-aware adjustment strategy, which calculates the relative height ratio between drone and satellite views, dynamically adjusting partition sizes to explicitly align semantic information between partition pairs. This strategy is integrated into a Scale-adaptive Local Partition Network (SaLPN), building upon an existing square partition strategy to extract both finegrained and global features. Additionally, we propose a saliencyguided refinement strategy to enhance part-level features, further improving retrieval accuracy. Extensive experiments validate that our height-aware, scale-adaptive approach achieves stateof-the-art geo-localization accuracy in various scale-inconsistent scenarios and exhibits strong robustness against scale variations. The code will be made publicly available.

cs.CV↗

arXiv2025

D&M: Enriching E-commerce Videos with Sound Effects by Key Moment Detection and SFX Matching

Videos showcasing specific products are increasingly important for E-commerce. Key moments naturally exist as the first appearance of a specific product, presentation of its distinctive features, the presence of a buying link, etc. Adding proper sound effects (SFX) to these key moments, or video decoration with SFX (VDSFX), is crucial for enhancing the user engaging experience. Previous studies about adding SFX to videos perform video to SFX matching at a holistic level, lacking the ability of adding SFX to a specific moment. Meanwhile, previous studies on video highlight detection or video moment retrieval consider only moment localization, leaving moment to SFX matching untouched. By contrast, we propose in this paper D&M, a unified method that accomplishes key moment detection and moment to SFX matching simultaneously. Moreover, for the new VDSFX task we build a large-scale dataset SFX-Moment from an E-commerce platform. For a fair comparison, we build competitive baselines by extending a number of current video moment detection methods to the new task. Extensive experiments on SFX-Moment show the superior performance of the proposed method over the baselines.

cs.CV↗

arXiv2025

A Language Vision Model Approach for Automated Tumor Contouring in Radiation Oncology

Background: Lung cancer ranks as the leading cause of cancer-related mortality worldwide. The complexity of tumor delineation, crucial for radiation therapy, requires expertise often unavailable in resource-limited settings. Artificial Intelligence(AI), particularly with advancements in deep learning (DL) and natural language processing (NLP), offers potential solutions yet is challenged by high false positive rates. Purpose: The Oncology Contouring Copilot (OCC) system is developed to leverage oncologist expertise for precise tumor contouring using textual descriptions, aiming to increase the efficiency of oncological workflows by combining the strengths of AI with human oversight. Methods: Our OCC system initially identifies nodule candidates from CT scans. Employing Language Vision Models (LVMs) like GPT-4V, OCC then effectively reduces false positives with clinical descriptive texts, merging textual and visual data to automate tumor delineation, designed to elevate the quality of oncology care by incorporating knowledge from experienced domain experts. Results: Deployments of the OCC system resulted in a significant reduction in the false discovery rate by 35.0%, a 72.4% decrease in false positives per scan, and an F1-score of 0.652 across our dataset for unbiased evaluation. Conclusions: OCC represents a significant advance in oncology care, particularly through the use of the latest LVMs to improve contouring results by (1) streamlining oncology treatment workflows by optimizing tumor delineation, reducing manual processes; (2) offering a scalable and intuitive framework to reduce false positives in radiotherapy planning using LVMs; (3) introducing novel medical language vision prompt techniques to minimize LVMs hallucinations with ablation study, and (4) conducting a comparative analysis of LVMs, highlighting their potential in addressing medical language vision challenges.

eess.IV↗

arXiv2025

Power Spectra of Velocity Fluctuations in Granular Heap Flow

This study used Speckle Visibility Spectroscopy to examine velocity fluctuations in a three-dimensional granular heap flow, where the mean velocity profile consists of a fast-flow surface layer and a creep layer beneath. The velocity spectra follow power-law scalings, $E(f) \propto f^α$, with $α\approx -0.85$ in the surface flow layer -- matching the Self-Organized Criticality (SOC) model with open boundaries ($α\approx -0.95$). In the creep layer, $α$ decreases with depth, reaching $α\approx -1.5$ at approximately 55 mean particle diameters, consistent with the SOC model with closed boundaries ($α\approx -1.58$). Analysis of the fluctuation velocity distributions offers additional insights into the microscopic origins of the spectrum's characteristics. These findings help resolve the long-standing puzzle of flow localization in gravity-driven granular flows despite a constant shear stress-to-pressure ratio throughout the material.

cond-mat.soft↗

arXiv2025

SweetTok: Semantic-Aware Spatial-Temporal Tokenizer for Compact Video Discretization

This paper presents the \textbf{S}emantic-a\textbf{W}ar\textbf{E} spatial-t\textbf{E}mporal \textbf{T}okenizer (SweetTok), a novel video tokenizer to overcome the limitations in current video tokenization methods for compacted yet effective discretization. Unlike previous approaches that process flattened local visual patches via direct discretization or adaptive query tokenization, SweetTok proposes a decoupling framework, compressing visual inputs through distinct spatial and temporal queries via \textbf{D}ecoupled \textbf{Q}uery \textbf{A}uto\textbf{E}ncoder (DQAE). This design allows SweetTok to efficiently compress video token count while achieving superior fidelity by capturing essential information across spatial and temporal dimensions. Furthermore, we design a \textbf{M}otion-enhanced \textbf{L}anguage \textbf{C}odebook (MLC) tailored for spatial and temporal compression to address the differences in semantic representation between appearance and motion information. SweetTok significantly improves video reconstruction results by \textbf{42.8\%} w.r.t rFVD on UCF-101 dataset. With a better token compression strategy, it also boosts downstream video generation results by \textbf{15.1\%} w.r.t gFVD. Additionally, the compressed decoupled tokens are imbued with semantic information, enabling few-shot recognition capabilities powered by LLMs in downstream applications.

cs.CV↗

arXiv2025

Efficient Function-as-a-Service for Large Language Models with TIDAL

Large Language Model (LLM) applications have emerged as a prominent use case for Function-as-a-Service (FaaS) due to their high computational demands and sporadic invocation patterns. However, serving LLM functions within FaaS frameworks faces significant GPU-side cold start. A fundamental approach involves leveraging a template with function state saved on GPUs to bypass the cold start for new invocations. Yet, this approach struggles with the high GPU footprint, dynamic initialization behaviors, and lazy GPU kernel loading inherent in LLM functions, primarily due to a lack of insight into the underlying execution details. In this paper, we introduce TIDAL, an optimized FaaS framework for LLM applications that achieves fast startups by tracing fine-grained execution paths. By utilizing the traced execution details, TIDAL generates adaptive function templates, effectively breaking startup barriers for LLM functions. Extensive evaluations demonstrate that TIDAL reduces cold start latency by $1.79\times\text{\textasciitilde}2.11\times$ and improves the $95\%$-ile time-to-first-token by $76.0\%$, surpassing state-of-the-art methods.

cs.OS↗

arXiv2025

Neural Network Surrogate Model for Junction Temperature and Hotspot Position in $3$D Multi-Layer High Bandwidth Memory (HBM) Chiplets under Varying Thermal Conditions

As the demand for computational power increases, high-bandwidth memory (HBM) has become a critical technology for next-generation computing systems. However, the widespread adoption of HBM presents significant thermal management challenges, particularly in multilayer through-silicon-via (TSV) stacked structures under varying thermal conditions, where accurate prediction of junction temperature and hotspot position is essential during the early design. This work develops a data-driven neural network model for the fast prediction of junction temperature and hotspot position in 3D HBM chiplets. The model, trained with a data set of $13,494$ different combinations of thermal condition parameters, sampled from a vast parameter space characterized by high-dimensional combination (up to $3^{27}$), can accurately and quickly infer the junction temperature and hotspot position for any thermal conditions in the parameter space. Moreover, it shows good generalizability for other thermal conditions not considered in the parameter space. The data set is constructed using accurate finite element solvers. This method not only minimizes the reliance on costly experimental tests and extensive computational resources for finite element analysis but also accelerates the design and optimization of complex HBM systems, making it a valuable tool for improving thermal management and performance in high-performance computing applications.

cs.LG↗

arXiv2025

Comet: Fine-grained Computation-communication Overlapping for Mixture-of-Experts

Mixture-of-experts (MoE) has been extensively employed to scale large language models to trillion-plus parameters while maintaining a fixed computational cost. The development of large MoE models in the distributed scenario encounters the problem of large communication overhead. The inter-device communication of a MoE layer can occupy 47% time of the entire model execution with popular models and frameworks. Therefore, existing methods suggest the communication in a MoE layer to be pipelined with the computation for overlapping. However, these coarse grained overlapping schemes introduce a notable impairment of computational efficiency and the latency concealing is sub-optimal. To this end, we present COMET, an optimized MoE system with fine-grained communication-computation overlapping. Leveraging data dependency analysis and task rescheduling, COMET achieves precise fine-grained overlapping of communication and computation. Through adaptive workload assignment, COMET effectively eliminates fine-grained communication bottlenecks and enhances its adaptability across various scenarios. Our evaluation shows that COMET accelerates the execution of a single MoE layer by $1.96\times$ and for end-to-end execution, COMET delivers a $1.71\times$ speedup on average. COMET has been adopted in the production environment of clusters with ten-thousand-scale of GPUs, achieving savings of millions of GPU hours.

cs.DC↗