arXiv subjects

Hui Guan

Publications and source records attributed to Hui Guan.

At least 19 recordsLinked to original sources

arXiv2026

Kalypso: Relational LLM Serving

Large language models are increasingly used as semantic operators for filtering, extracting, ranking, joining, and transforming unstructured data. Existing semantic query processing systems invoke request-centric LLM serving systems that are unaware of the query plan, leaving substantial performance opportunities unused. This paper introduces relational LLM serving, an abstraction that makes LLM serving aware of semantic query structure while preserving query semantics and output accuracy. The key opportunity is pipelined execution across semantic operators: when intermediate tuples flow directly from one operator to the next, their KV-cache state can be reused instead of recomputed. We present Kalypso, a relational LLM serving system that exposes an API for semantic query plans and executes them using an adaptive, memory-aware scheduling algorithm. Kalypso addresses a new online scheduling problem in which pipelined operator execution is coupled with GPU memory pressure management to reuse KV-cache state in the serving engine before eviction. Its scheduler continuously adjusts memory allocations to balance upstream parallelism, downstream progress, and GPU utilization. Our evaluation shows that Kalypso improves query completion time over baselines using request-centric LLM serving, with speedups up to 4.57x across diverse workloads, demonstrating that query-aware LLM serving can substantially improve the efficiency of semantic query execution.

cs.DB↗

arXiv2026

Mosaic: Runtime-Efficient Multi-Agent Embodied Planning

LLM-based multi-agent embodied planning remains impractical due to prohibitively high execution latency. We identify failed actions as the dominant bottleneck, stemming from two core challenges: inaccurate state tracking under partial observability and inefficient coordination that produces redundant or conflicting actions. We introduce Mosaic, a runtime-efficient multi-agent planning framework that addresses both challenges. Mosaic maintains accurate yet lightweight state tracking through agent-centric semantic memory that stores objects in relative coordinates, enabling geometric transformations and coordination. It ensures efficient coordination through Integer Linear Programming that allocates actions at every planning step, enforcing physical feasibility and inter-agent coordination constraints. Across AI2-THOR and search-and-rescue benchmarks, Mosaic achieves 27-32% faster execution, 30-33% fewer LLM calls, 25-31% fewer steps, and 4-10% points higher success rates. These results demonstrate that efficient memory and constraint-guided coordination are critical for scalable, low-latency multi-agent planning.

cs.MA↗

arXiv2026

EventADL: Open-Box Anomaly Detection and Localization Framework for Events in Cloud-Based Service Systems

Anomaly detection and localization (ADL) is critical for maintaining reliability and availability in cloud systems. Recent ADL developments focus on metric and log data, leaving event data unexplored. To address this gap, we propose EventADL, the first open-box event-based ADL framework for cloud-based service systems. To motivate the design of our framework, we conduct a systematic analysis on 520 real-world incidents, and provide insights into how anomalies and their root causes manifest through event data. EventADL has three phases: offline training, online anomaly detection, and root cause localization. During the training phase, EventADL first learns Event Semantic Patterns (ESPs), which capture normal interactions between system entities using historical event data, and then learns Event Frequency Patterns (EFPs), which capture the normal frequency of known ESPs. In the online anomaly detection phase, any data in the event stream that deviates significantly from either pattern is identified as anomalous. For localization, EventADL constructs an Intervention Graph that models the relationships between recent system interactions and the detected anomalies for automatic root cause localization. The framework is designed to operate efficiently with unlabeled data and to produce interpretable anomalies with their corresponding root causes. Our evaluation on three real cloud service systems and two real-world incidents demonstrates that EventADL outperforms existing methods, achieving F1-scores of at least 90% for anomaly detection and 100% top-3 accuracy in root cause localization.

cs.LG↗

arXiv2026

Transferable FB-GNN-MBE Framework for Potential Energy Surfaces: Data-Adaptive Transfer Learning in Deep Learned Many-Body Expansion Theory

Mechanistic understanding and rational design of complex chemical systems depend on fast and accurate predictions of electronic structures beyond individual building blocks. However, if the system exceeds hundreds of atoms, first-principles quantum mechanical (QM) modeling becomes impractical. In this study, we developed FB-GNN-MBE by integrating a fragment-based graph neural network (FB-GNN) into the many-body expansion (MBE) theory and demonstrated its capacity to reproduce first-principles potential energy surfaces (PES) for hierarchically structured systems with manageable accuracy, complexity, and interpretability. Specifically, we divided the entire system into basic building blocks (fragments), evaluated their one-fragment energies using a QM model, and addressed many-fragment interactions using the structure-property relationships trained by FB-GNNs. Our investigation shows that FB-GNN-MBE achieves chemical accuracy in predicting two-body (2B) and three-body (3B) energies across water, phenol, and mixture benchmarks, as well as the one-dimensional dissociation curves of water and phenol dimers. To transfer the success of FB-GNN-MBE across various systems with minimal computational costs and data demands, we developed and validated a teacher-student learning protocol. A heavy-weight FB-GNN trained on a mixed-density water cluster ensemble (teacher) distills its learned knowledge and passes it to a light-weight GNN (student), which is later fine-tuned on a uniform-density (H2O)21 cluster ensemble. This transfer learning strategy resulted in efficient and accurate prediction of 2B and 3B energies for variously sized water clusters without retraining. Our transferable FB-GNN-MBE framework outperformed conventional non-FB-GNN-based models and provided a scalable and accurate route toward interaction energies of large molecular assemblies.

physics.chem-ph↗

arXiv2026

Interactive Clarification for Cloud Infrastructure-as-Code Synthesis

The scale and complexity of modern cloud infrastructure have made "Infrastructure-as-Code" (IaC) essential for managing deployments through declarative configurations. While large language models (LLMs) are increasingly used to generate IaC configurations from natural language, user requests are often ambiguous and underspecified. Unlike traditional code generation, it is costly and time-consuming to test IaC configurations during the synthesis, forcing the LLMs into an almost one-shot regime. We observe that ambiguity in IaC synthesis exhibits a compositional structure: configurations decompose into three axes (resources, topology, attributes) where higher-level decisions constrain lower-level ones. We propose a training-free, multi-level disambiguation framework that generates diverse candidate specifications, identifies structural disagreements across these axes, ranks them by informativeness, and produces targeted clarification questions that progressively narrow the configuration space. We further introduce Ambig-IaC, an expert-verified benchmark of 300 validated IaC tasks with ambiguous requests, and define evaluation metrics based on graph edit distance and exact attribute matching. Comprehensive experiments show that our method outperforms existing interactive clarification baselines, with gains that scale with the interaction budget and generalize across models. Extensive ablation studies and analyses further demonstrate its robustness for interactive IaC synthesis.

cs.SE↗

arXiv2025

Atom: Efficient On-Device Video-Language Pipelines Through Modular Reuse

Recent advances in video-language models have enabled powerful applications like video retrieval, captioning, and assembly. However, executing such multi-stage pipelines efficiently on mobile devices remains challenging due to redundant model loads and fragmented execution. We introduce Atom, an on-device system that restructures video-language pipelines for fast and efficient execution. Atom decomposes a billion-parameter model into reusable modules, such as the visual encoder and language decoder, and reuses them across subtasks like captioning, reasoning, and indexing. This reuse-centric design eliminates repeated model loading and enables parallel execution, reducing end-to-end latency without sacrificing performance. On commodity smartphones, Atom achieves 27--33% faster execution compared to non-reuse baselines, with only marginal performance drop ($\leq$ 2.3 Recall@1 in retrieval, $\leq$ 1.5 CIDEr in captioning). These results position Atom as a practical, scalable approach for efficient video-language understanding on edge devices.

cs.LG↗

arXiv2025

Automated Cloud Infrastructure-as-Code Reconciliation with AI Agents

Cloud infrastructure is managed through a mix of interfaces -- traditionally, cloud consoles, command-line interfaces (CLI), and SDKs are the tools of choice. Recently, Infrastructure-as-Code/IaC frameworks (e.g., Terraform) have quickly gained popularity. Unlike conventional tools, IaC~frameworks encode the infrastructure in a "source-of-truth" configuration. They are capable of automatically carrying out modifications to the cloud -- deploying, updating, or destroying resources -- to bring the actual infrastructure into alignment with the IaC configuration. However, when IaC is used alongside consoles, CLIs, or SDKs, it loses visibility into external changes, causing infrastructure drift, where the configuration becomes outdated, and later IaC operations may undo valid updates or trigger errors. We present NSync, an automated system for IaC reconciliation that propagates out-of-band changes back into the IaC program. Our key insight is that infrastructure changes eventually all occur via cloud API invocations -- the lowest layer for cloud management operations. NSync gleans insights from API traces to detect drift (i.e., non-IaC changes) and reconcile it (i.e., update the IaC configuration to capture the changes). It employs an agentic architecture that leverages LLMs to infer high-level intents from noisy API sequences, synthesize targeted IaC updates using specialized tools, and continually improve through a self-evolving knowledge base of past reconciliations. We further introduce a novel evaluation pipeline for injecting realistic drifts into cloud infrastructure and assessing reconciliation performance. Experiments across five real-world Terraform projects and 372 drift scenarios show that NSync outperforms the baseline both in terms of accuracy (from 0.71 to 0.97 pass@3) and token efficiency (1.47$\times$ improvement).

cs.SE↗

arXiv2025

HADIS: Hybrid Adaptive Diffusion Model Serving for Efficient Text-to-Image Generation

Text-to-image diffusion models have achieved remarkable visual quality but incur high computational costs, making latency-aware, scalable deployment challenging. To address this, we advocate a hybrid architecture that achieves query awareness when serving diffusion models. Unlike existing query-aware serving systems that cascade lightweight and heavyweight models with a fixed configuration, our hybrid architecture first routes each query directly to a suitable model variant, then reroutes it to a cascaded heavyweight model only if necessary. We theoretically analyze conditions for the hybrid architecture to outperform non-hybrid alternatives in latency and response quality. Building on this architecture, we design HADIS, a hybrid serving system for latency-aware diffusion models that jointly optimizes cascade model selection, query routing, and resource allocation. To reduce the complexity of resource management, HADIS uses an offline profiling phase to produce a Pareto-optimal cascade configuration table. At runtime, HADIS selects the best cascade configuration and GPU allocation given latency and workload constraints. Empirical evaluations on real-world traces demonstrate that HADIS improves response quality by up to 35% while reducing latency violation rates by 2.7-45$\times$ compared to state-of-the-art model serving systems.

cs.DC↗

arXiv2025

Adaptive Spiking with Plasticity for Energy Aware Neuromorphic Systems

This paper presents ASPEN, a novel energy-aware technique for neuromorphic systems that could unleash the future of intelligent, always-on, ultra-low-power, and low-burden wearables. Our main research objectives are to explore the feasibility of neuromorphic computing for wearables, identify open research directions, and demonstrate the feasibility of developing an adaptive spiking technique for energy-aware computation, which can be game-changing for resource-constrained devices in always-on applications. As neuromorphic computing systems operate based on spike events, their energy consumption is closely related to spiking activity, i.e., each spike incurs computational and power costs; consequently, minimizing the number of spikes is a critical strategy for operating under constrained energy budgets. To support this goal, ASPEN utilizes stochastic perturbations to the neuronal threshold during training to not only enhance the network's robustness across varying thresholds, which can be controlled at inference time, but also act as a regularizer that improves generalization, reduces spiking activity, and enables energy control without the need for complex retraining or pruning. More specifically, ASPEN adaptively adjusts intrinsic neuronal parameters as a lightweight and scalable technique for dynamic energy control without reconfiguring the entire model. Our evaluation on neuromorphic emulator and hardware shows that ASPEN significantly reduces spike counts and energy consumption while maintaining accuracy comparable to state-of-the-art methods.

cs.NE↗

arXiv2025

Memory Savings at What Cost? A Study of Alternatives to Backpropagation

Forward-mode automatic differentiation (FmAD) and zero-order (ZO) optimization are increasingly proposed as memory-efficient, backpropagation-free alternatives for large language model (LLM) fine-tuning, yet their benefits are typically evaluated only against standard backpropagation (BP), omitting memory-efficient variants such as activation checkpointing. We present a unified theoretical and empirical comparison of BP, checkpointed BP, FmAD, and ZO for LLM and vision-language model training, showing that while FmAD and ZO reduce activation memory, they trade memory for higher computational cost and longer wall-clock time to convergence, resulting in lower accuracy and slower training, especially under constrained perturbation budgets. Across models, BP with checkpointing outperforms FmAD and ZO variants, including variance-reduced methods, achieving up to 31.1% higher accuracy, 34.8% faster convergence, and 3.8x fewer computations at comparable memory usage, while also revealing instability-related failure modes in FmAD and ZO. Overall, our results correct a one-sided benchmarking narrative by showing that memory-efficient methods entail fundamentally different trade-offs, and that ignoring these distinctions has led to misleading conclusions about LLM optimization in prior work. Our source code is available at {https://github.com/Astuary/Gradient_Estimation_Methods}.

cs.LG↗

arXiv2025

Reimagining Parameter Space Exploration with Diffusion Models

Adapting neural networks to new tasks typically requires task-specific fine-tuning, which is time-consuming and reliant on labeled data. We explore a generative alternative that produces task-specific parameters directly from task identity, eliminating the need for task-specific training. To this end, we propose using diffusion models to learn the underlying structure of effective task-specific parameter space and synthesize parameters on demand. Once trained, the task-conditioned diffusion model can generate specialized weights directly from task identifiers. We evaluate this approach across three scenarios: generating parameters for a single seen task, for multiple seen tasks, and for entirely unseen tasks. Experiments show that diffusion models can generate accurate task-specific parameters and support multi-task interpolation when parameter subspaces are well-structured, but fail to generalize to unseen tasks, highlighting both the potential and limitations of this generative solution.

cs.LG↗

arXiv2025

ASTRA: Communication-Efficient Acceleration for Multi-Device Transformer Inference

Multi-device inference can reduce Transformer latency by parallelizing computation. However, existing methods require high inter-device bandwidth, making them impractical for bandwidth-constrained environments. We present ASTRA, a communication-efficient framework that integrates sequence parallelism with mixed-precision attention, where non-local token embeddings are transmitted as low-bit vector-quantized codes while local attention remains full precision. To preserve accuracy under aggressive compression, ASTRA introduces Noise-Augmented Quantization and Distributed Class Tokens. Across vision and language models (e.g., ViT and GPT2), ASTRA achieves up to 2.64$\times$ speedup over single-device inference and up to 15.25$\times$ over prior multi-device baselines while operating at bandwidths as low as 10 Mbps. ASTRA remains robust on large models (e.g., Llama-3-8B) even under non-ideal network conditions such as packet loss and dynamic networks.

cs.LG↗

arXiv2024

AI-Driven Innovations in Volumetric Video Streaming: A Review

Recent efforts to enhance immersive and interactive user experiences have driven the development of volumetric video, a form of 3D content that enables 6 DoF. Unlike traditional 2D content, volumetric content can be represented in various ways, such as point clouds, meshes, or neural representations. However, due to its complex structure and large amounts of data size, deploying this new form of 3D data presents significant challenges in transmission and rendering. These challenges have hindered the widespread adoption of volumetric video in daily applications. In recent years, researchers have proposed various AI-driven techniques to address these challenges and improve the efficiency and quality of volumetric content streaming. This paper provides a comprehensive overview of recent advances in AI-driven approaches to facilitate volumetric content streaming. Through this review, we aim to offer insights into the current state-of-the-art and suggest potential future directions for advancing the deployment of volumetric video streaming in real-world applications.

cs.CV↗

arXiv2024

DiffServe: Efficiently Serving Text-to-Image Diffusion Models with Query-Aware Model Scaling

Text-to-image generation using diffusion models has gained increasing popularity due to their ability to produce high-quality, realistic images based on text prompts. However, efficiently serving these models is challenging due to their computation-intensive nature and the variation in query demands. In this paper, we aim to address both problems simultaneously through query-aware model scaling. The core idea is to construct model cascades so that easy queries can be processed by more lightweight diffusion models without compromising image generation quality. Based on this concept, we develop an end-to-end text-to-image diffusion model serving system, DiffServe, which automatically constructs model cascades from available diffusion model variants and allocates resources dynamically in response to demand fluctuations. Our empirical evaluations demonstrate that DiffServe achieves up to 24% improvement in response quality while maintaining 19-70% lower latency violation rates compared to state-of-the-art model serving systems.

cs.DC↗

arXiv2024

Aligned Vector Quantization for Edge-Cloud Collabrative Vision-Language Models

Vision Language Models (VLMs) are central to Visual Question Answering (VQA) systems and are typically deployed in the cloud due to their high computational demands. However, this cloud-only approach underutilizes edge computational resources and requires significant bandwidth for transmitting raw images. In this paper, we introduce an edge-cloud collaborative VQA system, called LLaVA-AlignedVQ, which features a novel Aligned Vector Quantization algorithm (AlignedVQ) that efficiently compress intermediate features without compromising accuracy to support partitioned execution. Our experiments demonstrate that LLaVA-AlignedVQ achieves approximately 1365x compression rate of intermediate features, reducing data transmission overhead by 96.8% compared to transmitting JPEG90-compressed images to the cloud. LLaVA-AlignedVQ achieves an inference speedup of 2-15x while maintaining high accuracy, remaining within -2.23% to +1.6% of the original model's accuracy performance across eight VQA datasets, compared to the cloud-only solution.

cs.CV↗

arXiv2024

Integrating Graph Neural Networks and Many-Body Expansion Theory for Potential Energy Surfaces

Rational design of next-generation functional materials relied on quantitative predictions of their electronic structures beyond single building blocks. First-principles quantum mechanical (QM) modeling became infeasible as the size of a material grew beyond hundreds of atoms. In this study, we developed a new computational tool integrating fragment-based graph neural networks (FBGNN) into the fragment-based many-body expansion (MBE) theory, referred to as FBGNN-MBE, and demonstrated its capacity to reproduce full-dimensional potential energy surfaces (FD-PES) for hierarchic chemical systems with manageable accuracy, complexity, and interpretability. In particular, we divided the entire system into basic building blocks (fragments), evaluated their single-fragment energies using a first-principles QM model and attacked many-fragment interactions using the structure-property relationships trained by FBGNNs. Our development of FBGNN-MBE demonstrated the potential of a new framework integrating deep learning models into fragment-based QM methods, and marked a significant step towards computationally aided design of large functional materials.

cond-mat.mtrl-sci↗

arXiv2024

Understanding and Alleviating Memory Consumption in RLHF for LLMs

Fine-tuning with Reinforcement Learning with Human Feedback (RLHF) is essential for aligning large language models (LLMs). However, RLHF often encounters significant memory challenges. This study is the first to examine memory usage in the RLHF context, exploring various memory management strategies and unveiling the reasons behind excessive memory consumption. Additionally, we introduce a simple yet effective approach that substantially reduces the memory required for RLHF fine-tuning.

cs.LG↗

arXiv2024

WildFit: Autonomous In-situ Model Adaptation for Resource-Constrained IoT Systems

Resource-constrained IoT devices increasingly rely on deep learning models, however, these models experience significant accuracy drops due to domain shifts when encountering variations in lighting, weather, and seasonal conditions. While cloud-based retraining can address this issue, many IoT deployments operate with limited connectivity and energy constraints, making traditional fine-tuning approaches impractical. We explore this challenge through the lens of wildlife ecology, where camera traps must maintain accurate species classification across changing seasons, weather, and habitats without reliable connectivity. We introduce WildFit, an autonomous in-situ adaptation framework that leverages the key insight that background scenes change more frequently than the visual characteristics of monitored species. WildFit combines background-aware synthesis to generate training samples on-device with drift-aware fine-tuning that triggers model updates only when necessary to conserve resources. Our background-aware synthesis surpasses efficient baselines by 7.3% and diffusion models by 3.0% while being orders of magnitude faster, our drift-aware fine-tuning achieves Pareto optimality with 50% fewer updates and 1.5% higher accuracy, and the end-to-end system outperforms domain adaptation approaches by 20-35% while consuming only 11.2 Wh over 37 days-enabling battery-powered deployment.

cs.CV↗