arXiv subjects

Rong Gu

Publications and source records attributed to Rong Gu.

At least 19 recordsLinked to original sources

arXiv2026

LOCAL: Enabling Learning On-device Contiguously for Agent LLMs

On-device LLM agents interact repeatedly with users on local hardware, producing private traces that are valuable for adaptation but should not be sent to a remote trainer. Ideally, such agents would learn contiguously---adapting from every interaction without pausing or suspending user-facing inference---yet existing inference runtimes assume stable weights and existing RL systems assume separated resources, so neither can support this continuity. We present LOCAL, the first single-GPU runtime that enables contiguous on-device learning for LLM agents. The key insight is that GPU scheduling, adapter version management, and KV-cache validity cannot be handled by independent subsystems: adapter updates invalidate cached KV tensors from older versions, and cache retention affects the memory available for training. LOCAL makes adapter version, task priority, and cache state visible to three cooperating components---a cooperative scheduler, a version-aware KV-cache manager, and a multi-agent model runtime---that share this state to keep scheduling, execution, and cache maintenance mutually consistent. On a single 24 GB GPU with 7B-class models, LOCAL lowers foreground queue-wait p95 by 3.1x over FIFO, lowers p95 time-to-first-token (TTFT) by 1.55x versus non-preemptible training, cuts post-publish first-hit prefill p99 by 25.6% and cross-agent TTFT p99 by 21.9%, and keeps background learning progressing under tight KV budgets.

cs.DC↗

arXiv2026

TIDE-MC: Two-Sided Interpolative Decomposition for Billion-Scale GPU Matrix Completion

Matrix completion supports large-scale recommendation and scientific computing, yet existing GPU solvers commonly assume that the observed matrix or its dense factors fit in device memory. On real workloads, this assumption leads to out-of-memory failures or severe PCIe overhead under naive paging. We present TIDE-MC, a bounded-memory GPU framework built on Two-Sided Interpolative Decomposition (TSID). TSID uses a sampled template submatrix as an anchor for reconstructing the full low-rank matrix, allowing computation and storage to scale with the template and active data chunks rather than the complete matrix. TIDE-MC realizes this formulation through two execution stages. First, a conflict-free synchronization engine recovers the template using parallel factorization and hierarchical gradient aggregation. Second, a chunked reconstruction pipeline extends the recovered template to the remaining matrix while overlapping PCIe transfers with GPU computation. An asymmetric gradient-clipping scheme stabilizes mixed-precision Tensor Core execution. Across 15 benchmarks, TIDE-MC completes workloads that cause existing GPU solvers to run out of memory. Compared with the evaluated state-of-the-art baselines, it achieves up to 11,647x speedup, reduces peak memory usage by up to 8.5x, and lowers reconstruction error by up to 99.7%. These results show that template-anchored decomposition and stage-specific GPU execution can scale matrix completion beyond device-memory capacity.

cs.DC↗

arXiv2026

SpecLA: Efficient Speculative Decoding for Linear-Attention Models

Linear-attention models replace the growing KV cache with recurrent states, but autoregressive decoding still reads, updates, and writes these states one token at a time. Speculative decoding can reduce this cost by verifying several draft tokens in one target pass, yet existing speculative systems are designed for Transformer KV caches. For stateful linear-attention targets, verification must follow recurrent dependencies across chains and branches, acceptance must update only the accepted state trajectory, and the drafter must avoid submitting candidates that waste stateful verification work. This paper presents SpecLA, a speculative decoding runtime for stateful linear-attention models. SpecLA verifies chains and trees with topology-aware kernels, stores compact factors produced during verification to recover accepted states, and uses confidence pruning plus a target-aligned EAGLE-style drafter to feed useful candidates to the verifier. On an NVIDIA H100 with a public GDN-1.3B target, SpecLA achieves up to 1.70x end-to-end speedup over autoregressive decoding.

cs.CL↗

arXiv2026

CacheWeaver: Cache-Aware Evidence Ordering for Efficient Grounded RAG Inference

Retrieval-Augmented Generation (RAG) improves factual grounding, but it also lengthens prompts and raises prefill cost. Prefix caching in serving engines such as vLLM reduces this cost only when requests share the same token prefix. In grounded generation, however, adjacent queries may retrieve overlapping evidence in different orders, so set overlap does not become reusable prefix overlap. We present CacheWeaver, a lightweight prompt-layer method for cache-aware evidence ordering. The method keeps a prefix tree over recently served evidence sequences and uses a greedy walk to place the most reusable prefix first, while leaving the serving engine and retrieved evidence set unchanged. Across three vLLM configurations, the method lowers median time-to-first-token (TTFT) by about 20-33 percent relative to retrieval-order prefix caching, without hurting answer quality in our QA tests. The greedy policy reaches 97.5 percent of the median TTFT gain from oracle ordering, indicating that most reusable prefix locality can be recovered by a simple scheduling layer between retrieval and inference.

cs.CL↗

arXiv2026

SSV: Sparse Speculative Verification for Efficient LLM Inference

Speculative decoding and dynamic sparse attention are two complementary approaches for accelerating long-context LLM inference: the former amortizes target-model execution across multiple verifier queries, while the latter reduces each query's KV-cache working set. Directly combining them, however, exposes a structural mismatch: speculative verification relies on cross-query commonality, whereas dynamic sparse attention assigns query-specific sparse layouts. This mismatch limits KV-block reuse, amplifies NSA's branch-wise overheads, and makes verification strategy selection input- and regime-dependent. We present SSV, a sparse speculative-verification framework that turns dynamic sparse attention into a verification-oriented workload. SSV combines overlap-aware grouped-query execution, refresh/reuse-based NSA kernel fusion, and profile-guided prompt-adaptive orchestration to improve cross-query reuse, reduce selected-index and branch-fusion overheads, and select effective draft-verification strategies under user-specified precision classes. Experiments on NVIDIA H100 GPUs show that SSV achieves up to 3.49x end-to-end throughput over autoregressive NSA decoding and up to 6.86x kernel speedups for sparse speculative verification.

cs.OS↗

arXiv2025

Scaling Graph Chain-of-Thought Reasoning: A Multi-Agent Framework with Efficient LLM Serving

Graph Chain-of-Thought (Graph-CoT) enables large language models (LLMs) to perform step-by-step reasoning over graph-structured knowledge, but existing pipelines suffer from low accuracy, excessive token usage, high latency, and low throughput due to single-agent monolithic prompts, repeated context re-encoding, and inefficient serving execution. We present GLM, the first multi-agent Graph-CoT system co-designed with an optimized LLM serving architecture. GLM decomposes reasoning into specialized agents for classification, reasoning, action generation, and graph retrieval, enabling branching and selective context sharing to reduce prompt length and reasoning iterations while preserving reasoning quality, thereby improving accuracy and reducing overall token consumption. To scale inference, we introduce a Graph-CoT-aware LLM inference mechanism with graph-specific KV-cache management, priority-based eviction, and pipelined execution to improve serving efficiency. Experiments demonstrate that GLM improves answer accuracy by up to 38%, reduces token cost by up to 95.7%, lowers inference latency by 90.3%, and achieves up to 15.1x higher throughput compared to state-of-the-art Graph-CoT baselines, enabling efficient adoption for complex real-world reasoning at scale.

cs.LG↗

arXiv2025

STAR: Decode-Phase Rescheduling for LLM Inference

Large Language Model (LLM) inference has emerged as a fundamental paradigm, however, variations in output length cause severe workload imbalance in the decode phase, particularly for long-output reasoning tasks. Existing systems, such as PD disaggregation architectures, rely on static prefill-to-decode scheduling, which often results in SLO violations and OOM failures under evolving decode workloads. In this paper, we propose STAR, a decode rescheduling system powered by length prediction to anticipate future workloads. Our core contributions include: (1) A lightweight and continuous LLM-native prediction method that leverages LLM hidden state to model remaining generation length with high precision (reducing MAE by 49.42%) and low overhead (cutting predictor parameters by 93.28%); (2) A rescheduling solution in decode phase with a dynamic balancing mechanism that integrates current and predicted workloads, reducing P99 TPOT by 75.1% and achieving 2.63 times higher goodput.

cs.DC↗

arXiv2025

SmartSwap: Swap-Based Memory Optimization for LLM Training under Varying Operator Sequences

The increasing size of large language models (LLMs) has led to a surge in memory requirements during training, often exceeding the capacity of high-bandwidth memory (HBM). Swap-based memory optimization incurs neither accuracy loss nor additional end-to-end overhead when effectively overlapped, thus being an attractive solution. However, existing swap methods assume consistent operator sequences, which is impractical in Eager Mode, where operator sequences can vary during change. We propose Chameleon, which redesigns the end-to-end process of swap-based memory optimization and is the first work to consider varying operator sequences in Eager Mode. Chameleon (i) introduces a lightweight online profiler to enable continuous profiling for monitoring operator sequences, (ii) generates effective swap policies with limited operator information, and (iii) optimizes the policy execution module for accurate policy application and better performance. Experimental results demonstrate that Chameleon reduces profiling overhead by 84.25%, enables training models up to 4x larger than hardware memory while adapting to changes in operator sequences, improves performance by up to 38.94% compared to recomputation or high-degree parallelism.

cs.DC↗

arXiv2025

Chameleon: Adaptive Fault Tolerance for Distributed Training via Real-time Policy Selection

Training large language models faces frequent interruptions due to various faults, demanding robust fault-tolerance. Existing backup-free methods, such as redundant computation, dynamic parallelism, and data rerouting, each incur performance penalties, whether from ongoing overhead, lengthy reconfigurations, or post-recovery inefficiencies. We propose Chameleon, an adaptive fault-tolerant system that intelligently selects optimal recovery strategies when a failure occurs. Chameleon achieves this through a unified performance model, expedient execution plan search, accurate performance estimation, and efficient communication optimizations. Experiments on a 32-card cluster show that Chameleon maintains a performance gap of within 11.00% between post-recovery and failure-free training, while preserving model convergence and efficient memory usage. Compared to state-of-the-art methods, Chameleon achieves up to 1.229x and 1.355x higher average throughput than Oobleck and Recycle, respectively.

cs.DC↗

arXiv2025

Accelerating Mixture-of-Experts Inference by Hiding Offloading Latency with Speculative Decoding

Recent advancements in Mixture of Experts (MoE) models have significantly increased their parameter scale as well as model performance. Extensive offloading techniques have been proposed to address the GPU memory limitations of MoE inference. However, due to the I/O bottleneck and sparse computation of MoE models, existing offloading techniques still suffer from low hardware utilization. To fully utilize the hardware resources, we propose SpecMoEOff, which employs the speculative decoding technique to enlarge the workload of each expert. SpecMoEOff orchestrates the GPU and CPU by both theoretical and empirical roofline analysis. In addition, we develop a dedicated CPU chunked attention verification kernel to fit the speculative decoding in offloading scenarios as well as minimizing the additional overhead led by draft models. SpecMoEOff further integrates an optimizer to automatically tune the hyperparameters of speculative decoding under given hardware and workload. Experimental results show that SpecMoEOff achieves up to 2.5x decode throughput improvement over the state-of-the-art MoE offloading techniques.

cs.DC↗

arXiv2025

Chordless Structure: A Pathway to Simple and Expressive GNNs

Researchers have proposed various methods of incorporating more structured information into the design of Graph Neural Networks (GNNs) to enhance their expressiveness. However, these methods are either computationally expensive or lacking in provable expressiveness. In this paper, we observe that the chords increase the complexity of the graph structure while contributing little useful information in many cases. In contrast, chordless structures are more efficient and effective for representing the graph. Therefore, when leveraging the information of cycles, we choose to omit the chords. Accordingly, we propose a Chordless Structure-based Graph Neural Network (CSGNN) and prove that its expressiveness is strictly more powerful than the k-hop GNN (KPGNN) with polynomial complexity. Experimental results on real-world datasets demonstrate that CSGNN outperforms existing GNNs across various graph tasks while incurring lower computational costs and achieving better performance than the GNNs of 3-WL expressiveness.

cs.LG↗

arXiv2025

CoDec: Prefix-Shared Decoding Kernel for LLMs

Prefix-sharing among multiple prompts presents opportunities to combine the operations of the shared prefix, while attention computation in the decode stage, which becomes a critical bottleneck with increasing context lengths, is a memory-intensive process requiring heavy memory access on the key-value (KV) cache of the prefixes. Therefore, in this paper, we explore the potential of prefix-sharing in the attention computation of the decode stage. However, the tree structure of the prefix-sharing mechanism presents significant challenges for attention computation in efficiently processing shared KV cache access patterns while managing complex dependencies and balancing irregular workloads. To address the above challenges, we propose a dedicated attention kernel to combine the memory access of shared prefixes in the decoding stage, namely CoDec. CoDec delivers two key innovations: a novel shared-prefix attention kernel that optimizes memory hierarchy and exploits both intra-block and inter-block parallelism, and a comprehensive workload balancing mechanism that efficiently estimates cost, divides tasks, and schedules execution. Experimental results show that CoDec achieves an average $1.9\times$ speedup and $120.9\times$ memory access reduction compared to the state-of-the-art FlashDecoding kernel regarding attention computation in the decode stage and $3.8\times$ end-to-end time per output token compared to the vLLM.

cs.LG↗

arXiv2025

Echo: Efficient Co-Scheduling of Hybrid Online-Offline Tasks for Large Language Model Serving

Large language models have been widely deployed in various applications, encompassing both interactive online tasks and batched offline tasks. Given the burstiness and latency sensitivity of online tasks, over-provisioning resources is common practice. This allows for the integration of latency-insensitive offline tasks during periods of low online load, enhancing resource utilization. However, strategically serving online and offline tasks through a preemption mechanism fails to fully leverage the flexibility of offline tasks and suffers from KV cache recomputation and irregular workloads. In this paper, we introduce Echo, a collaborative online-offline task serving system, including a scheduler, a KV cache manager, and estimation toolkits. The scheduler and KV cache manager work tightly to maximize the throughput of offline tasks, while the estimator further predicts execution time to ensure online task SLOs. The scheduler leverages the batch information of last iteration to reduce the search space for finding the optimal schedule. The KV cache manager sets the priority of the KV cache based on the type of tasks and the opportunity of prefix sharing to reduce the recomputation. Finally, the estimation toolkits predict the execution time, future memory consumption, and the throughput of offline tasks to guide the scheduler, KV cache manager, and the system deployer. Evaluation based on real-world workloads demonstrates that Echo can increase offline task throughput by up to $3.3\times$, while satisfying online task SLOs.

cs.DC↗

arXiv2025

A Composable Multimodal Framework for cine CMR-Text-Driven Prediction of Heart Failure Outcomes

Objective. Heart failure is one of the leading causes of death worldwide, with millions of deaths each year, according to data from the World Health Organization (WHO) and other public health agencies. While significant progress has been made in the field of heart failure, leading to improved survival rates and improvement of ejection fraction, there remains substantial unmet needs, due to the complexity and multifactorial characteristics. This study aims to propose and evaluate a composable strategy framework for assessment and treatment optimization in heart failure, designed to provide more holistic patient evaluation and management. Approach. The framework leverages multi-modal algorithms to analyze a comprehensive range of patient data, explicitly integrating cine cardiac magnetic resonance (cine CMR) sequences, structured clinical metrics (e.g., lab results, demographics), and unstructured textual records (e.g., medical history, prescriptions). By integrating these various data sources, our framework offers a more holistic evaluation and optimized treatment plan for patients. Main results. The multi-modal framework demonstrates superior accuracy in HF prognosis prediction compared to single-modal AI algorithms. Additionally, it enables a detailed evaluation of the impact of various pathological indicators on HF outcomes. Significance. By integrating heterogeneous clinical data in a systematic manner, this approach supports more comprehensive prognosis assessment and facilitates optimized, personalized treatment planning for heart failure patients.

cs.LG↗

arXiv2024

Model Checking for Reinforcement Learning in Autonomous Driving: One Can Do More Than You Think!

Most reinforcement learning (RL) platforms use high-level programming languages, such as OpenAI Gymnasium using Python. These frameworks provide various API and benchmarks for testing RL algorithms in different domains, such as autonomous driving (AD) and robotics. These platforms often emphasise the design of RL algorithms and the training performance but neglect the correctness of models and reward functions, which can be crucial for the successful application of RL. This paper proposes using formal methods to model AD systems and demonstrates how model checking (MC) can be used in RL for AD. Most studies combining MC and RL focus on safety, such as safety shields. However, this paper shows different facets where MC can strengthen RL. First, an MC-based model pre-analysis can reveal bugs with respect to sensor accuracy and learning step size. This step serves as a preparation of RL, which saves time if bugs exist and deepens users' understanding of the target system. Second, reward automata can benefit the design of reward functions and greatly improve learning performance especially when the learning objectives are multiple. All these findings are supported by experiments.

cs.LG↗

arXiv2024

Revisiting Service Level Objectives and System Level Metrics in Large Language Model Serving

User experience is a critical factor Large Language Model (LLM) serving systems must consider, where service level objectives (SLOs) considering the experience of individual requests and system level metrics (SLMs) considering the overall system performance are two key performance measures. However, we observe two notable issues in existing metrics: 1) manually delaying the delivery of some tokens can improve SLOs, and 2) actively abandoning requests that do not meet SLOs can improve SLMs, both of which are counterintuitive. In this paper, we revisit SLOs and SLMs in LLM serving, and propose a new SLO that aligns with user experience. Based on the SLO, we propose a comprehensive metric framework called smooth goodput, which integrates SLOs and SLMs to reflect the nature of user experience in LLM serving. Through this unified framework, we reassess the performance of different LLM serving systems under multiple workloads. Evaluation results show that our metric framework provides a more comprehensive view of token delivery and request processing, and effectively captures the optimal point of user experience and system performance with different serving strategies.

cs.LG↗

arXiv2024

CommonUppRoad: A Framework of Formal Modelling, Verifying, Learning, and Visualisation of Autonomous Vehicles

Combining machine learning and formal methods (FMs) provides a possible solution to overcome the safety issue of autonomous driving (AD) vehicles. However, there are gaps to be bridged before this combination becomes practically applicable and useful. In an attempt to facilitate researchers in both FMs and AD areas, this paper proposes a framework that combines two well-known tools, namely CommonRoad and UPPAAL. On the one hand, CommonRoad can be enhanced by the rigorous semantics of models in UPPAAL, which enables a systematic and comprehensive understanding of the AD system's behaviour and thus strengthens the safety of the system. On the other hand, controllers synthesised by UPPAAL can be visualised by CommonRoad in real-world road networks, which facilitates AD vehicle designers greatly adopting formal models in system design. In this framework, we provide automatic model conversions between CommonRoad and UPPAAL. Therefore, users only need to program in Python and the framework takes care of the formal models, learning, and verification in the backend. We perform experiments to demonstrate the applicability of our framework in various AD scenarios, discuss the advantages of solving motion planning in our framework, and show the scalability limit and possible solutions.

cs.MA↗

arXiv2021

Hash Adaptive Bloom Filter

Bloom filter is a compact memory-efficient probabilistic data structure supporting membership testing, i.e., to check whether an element is in a given set. However, as Bloom filter maps each element with uniformly random hash functions, few flexibilities are provided even if the information of negative keys (elements are not in the set) are available. The problem gets worse when the misidentification of negative keys brings different costs. To address the above problems, we propose a new Hash Adaptive Bloom Filter (HABF) that supports the customization of hash functions for keys. The key idea of HABF is to customize the hash functions for positive keys (elements are in the set) to avoid negative keys with high cost, and pack customized hash functions into a lightweight data structure named HashExpressor. Then, given an element at query time, HABF follows a two-round pattern to check whether the element is in the set. Further, we theoretically analyze the performance of HABF and bound the expected false positive rate. We conduct extensive experiments on representative datasets, and the results show that HABF outperforms the standard Bloom filter and its cutting-edge variants on the whole in terms of accuracy, construction time, query time, and memory space consumption (Note that source codes are available in [1]).

cs.DB↗