arXiv subjects

Xianzhong Ding

Publications and source records attributed to Xianzhong Ding.

16 recordsLinked to original sources

arXiv2026

AutoUVM: Automated Prefetching Framework for LLMs under UVM Oversubscription

Large language models (LLMs) increasingly exceed the memory capacity of commodity GPUs, making memory oversubscription common in practical deployments. NVIDIA Unified Virtual Memory (UVM) provides transparent access to host memory, but its page-fault-driven migrations introduce severe performance overhead. While UVM exposes primitives (e.g., prefetching and placement hints) to mitigate these costs, they require low-level CUDA modifications, limiting their applicability for most LLM users. Meanwhile, existing UVM optimizations operate at coarse managed-object granularity and fail to capture deep learning frameworks' internal tensor-level memory behavior, leading to excessive data movement and CPU-GPU interconnect bottlenecks. We propose AutoUVM, an automated, framework-aware UVM prefetching system for efficient LLM execution under memory oversubscription. AutoUVM bridges the semantic gap between deep learning frameworks and UVM by exposing tensor-level access information and enabling policy-driven prefetching at fine granularity. Implemented as a transparent extension, AutoUVM requires no changes to model code and dynamically adapts to runtime memory pressure. We instantiate AutoUVM with a roofline-inspired policy to identify performance-critical data transfers. Across ten LLMs, AutoUVM achieves an average 3.1x speedup over baseline UVM and consistently surpasses the best-performing prior UVM prefetcher by 1.9x, with improvements of up to 4.7x over object-level prefetchers, while significantly reducing page faults.

cs.OS↗

arXiv2026

ContextEcho: A Benchmark for Persona Drift in Long Agentic-Coding Sessions

A frontier language model's acknowledged "helpful programming assistant" persona does not survive long agentic-coding sessions in the deployment regime that production products actually run. After hours of tool-using debugging, a model that initially hedges preferences ("I don't have preferences") may begin asserting them ("Python - the feedback loop is instant..."), revealing user-visible drift that deployer evaluations may miss. Existing persona-stability studies focus on short dialogues and report little shift, leaving real-world code-generation regimes - thousands of tool-using turns, compaction, and hours-long sessions - largely uncharacterized. We introduce ContextEcho, a benchmark and reusable harness for measuring persona drift at deployment scale. It combines a 25-probe identity suite, a snapshot-then-probe protocol that forks conversation state without perturbing the main session, complementary judged and judge-free measurement surfaces, and three anonymized Claude Code sessions spanning 3,746-9,716 turns. Across 23 frontier models, ContextEcho shows that persona drift is general across organizations rather than family-specific, that in-session compaction does not reliably reset it, and that a single-shot anchor restores the trained register across measured targets. It also reveals mode-dependent downstream effects: while drift can facilitate tool-using continuation, in tool-free chat it breaks formatting contracts and inflates output length. Overall, ContextEcho provides researchers and deployers an open-source framework to audit whether the persona a model ships with is the persona users encounter at session end, across chat-completions API targets and without retraining.

cs.CL↗

arXiv2025

Optimizing Communication in Byzantine Agreement Protocols with Slim-HBBFT

Byzantine agreement protocols in asynchronous networks have received renewed interest because they do not rely on network behavior to achieve termination. Conventional asynchronous Byzantine agreement protocols require every party to broadcast its requests (e.g., transactions), and at the end of the protocol, parties agree on one party's request. If parties agree on one party's requests while exchanging every party's request, the protocol becomes expensive. These protocols are used to design an atomic broadcast (ABC) protocol where parties agree on $\langle n-f \rangle$ parties' requests (assuming $n=3f+1$, where $n$ is the total number of parties, and $f$ is the number of Byzantine parties). Although the parties agree on a subset of requests in the ABC protocol, if the requests do not vary (are duplicated), investing in a costly protocol is not justified. We propose Slim-HBBFT, an atomic broadcast protocol that considers requests from a fraction of $n$ parties and improves communication complexity by a factor of $O(n)$. At the core of our design is a prioritized provable-broadcast (P-PB) protocol that generates proof of broadcast only for selected parties. We use the P-PB protocol to design the Slim-HBBFT atomic broadcast protocol. Additionally, we conduct a comprehensive security analysis to demonstrate that Slim-HBBFT satisfies the properties of the Asynchronous Common Subset protocol, ensuring robust security and reliability.

cs.DC↗

arXiv2025

Prioritized-MVBA: A New Approach to Design an Optimal Asynchronous Byzantine Agreement Protocol

The multi-valued byzantine agreement protocol (MVBA) in the authenticated setting has been widely used as a core to design atomic broadcast and fault-tolerant state machine replication protocols in asynchronous networks. Originating from the seminal work of Cachin et al. \cite{CACHIN01}, subsequent research endeavors have sought to optimize protocol efficiency in terms of communication complexity. Notable advancements following Cachin's contributions include: i) VABA \cite{BYZ17}, requiring multiple protocol instances to achieve agreement on a party's request, and ii) Dumbo-MVBA \cite{LU20}, employing a cryptographic asynchronous dispersal and recovery methods to manage communication complexity alongside additional computational and communication rounds overheads. Our objective is to devise an MVBA protocol that achieves agreement in each instance without extra computation and communication rounds while maintaining the optimal metrics. Central to our design approach is the introduction of the committee in the classic MVBA protocol, wherein a randomly selected subset of ($f+1$, where $n=3f+1$) parties get selected and simultaneously broadcast their requests (transactions) to gather verifiable proofs. Successive distributions of these proofs afford us the necessary properties to employ the asynchronous binary Byzantine agreement (ABBA) protocol for reaching an agreement on a selected party's requests. By integrating the committee and ABBA protocols, we devise the optimal MVBA protocol, termed pMVBA (Prioritized-MVBA). This protocol exhibits resilience to tolerate up to $\lfloor \frac{n}{3}\rfloor$ Byzantine failures, with an expected runtime of $O(1)$, optimal message complexity of $O(n^2)$, and optimal communication complexity $O((l+λ)n^2)$ .

cs.DC↗

arXiv2025

Towards VM Rescheduling Optimization Through Deep Reinforcement Learning

Modern industry-scale data centers need to manage a large number of virtual machines (VMs). Due to the continual creation and release of VMs, many small resource fragments are scattered across physical machines (PMs). To handle these fragments, data centers periodically reschedule some VMs to alternative PMs, a practice commonly referred to as VM rescheduling. Despite the increasing importance of VM rescheduling as data centers grow in size, the problem remains understudied. We first show that, unlike most combinatorial optimization tasks, the inference time of VM rescheduling algorithms significantly influences their performance, due to dynamic VM state changes during this period. This causes existing methods to scale poorly. Therefore, we develop a reinforcement learning system for VM rescheduling, VM2RL, which incorporates a set of customized techniques, such as a two-stage framework that accommodates diverse constraints and workload conditions, a feature extraction module that captures relational information specific to rescheduling, as well as a risk-seeking evaluation enabling users to optimize the trade-off between latency and accuracy. We conduct extensive experiments with data from an industry-scale data center. Our results show that VM2RL can achieve a performance comparable to the optimal solution but with a running time of seconds. Code and datasets are open-sourced: https://github.com/zhykoties/VMR2L_eurosys, https://drive.google.com/drive/folders/1PfRo1cVwuhH30XhsE2Np3xqJn2GpX5qy.

cs.LG↗

arXiv2024

A Committee Based Optimal Asynchronous Byzantine Agreement Protocol W.P. 1

Multi-valued Byzantine agreement (MVBA) protocols are essential for atomic broadcast and fault-tolerant state machine replication in asynchronous networks. Despite advances, challenges persist in optimizing these protocols for communication and computation efficiency. This paper presents a committee-based MVBA protocol (cMVBA), a novel approach that achieves agreement without extra communication rounds by analyzing message patterns in asynchronous networks with probability 1.

cs.DC↗

arXiv2024

A Safe and Data-efficient Model-based Reinforcement Learning System for HVAC Control

Model-Based Reinforcement Learning (MBRL) has been widely studied for Heating, Ventilation, and Air Conditioning (HVAC) control in buildings. One of the critical challenges is the large amount of data required to effectively train neural networks for modeling building dynamics. This paper presents CLUE, an MBRL system for HVAC control in buildings. CLUE optimizes HVAC operations by integrating a Gaussian Process (GP) model to model building dynamics with uncertainty awareness. CLUE utilizes GP to predict state transitions as Gaussian distributions, effectively capturing prediction uncertainty and enhancing decision-making under sparse data conditions. Our approach employs a meta-kernel learning technique to efficiently set GP kernel hyperparameters using domain knowledge from diverse buildings. This drastically reduces the data requirements typically associated with GP models in HVAC applications. Additionally, CLUE incorporates these uncertainty estimates into a Model Predictive Path Integral (MPPI) algorithm, enabling the selection of safe, energy-efficient control actions. This uncertainty-aware control strategy evaluates and selects action trajectories based on their predicted impact on energy consumption and human comfort, optimizing operations even under uncertain conditions. Extensive simulations in a five-zone office building demonstrate that CLUE reduces the required training data from hundreds of days to just seven while maintaining robust control performance. It reduces comfort violations by an average of 12.07% compared to existing MBRL methods, without compromising on energy efficiency.

eess.SY↗

arXiv2024

Slim-ABC: An Optimized Atomic Broadcast Protocol

The Byzantine Agreement (BA) problem is a fundamental challenge in distributed systems, focusing on achieving reaching an agreement among parties, some of which may behave maliciously. With the rise of cryptocurrencies, there has been significant interest in developing atomic broadcast protocols, which facilitate agreement on a subset of parties' requests. However, these protocols often come with high communication complexity ($O(ln^2 + λn^3 \log n)$, where $l$ is the bit length of the input, $n$ is the number of parties, and $λ$ represents the security parameter bit length). This can lead to inefficiency, especially when the requests across parties exhibit little variation, resulting in unnecessary resource consumption. In this paper, we introduce Slim-ABC, a novel atomic broadcast protocol that eliminates the $O(ln^2 + λn^3 \log n)$ term associated with traditional atomic broadcast protocols. While Slim-ABC reduces the number of accepted requests, it significantly mitigates resource wastage, making it more efficient. The protocol leverages the asynchronous common subset and provable-broadcast mechanisms to achieve a communication complexity of $O(ln^2 + λn^2)$. Despite the trade-off in accepted requests, Slim-ABC maintains robust security by allowing only a fraction ($f+1$) of parties to broadcast requests. We present an extensive efficiency analysis of Slim-ABC, evaluating its performance across key metrics such as message complexity, communication complexity, and time complexity. Additionally, we provide a rigorous security analysis, demonstrating that Slim-ABC satisfies the \textit{agreement}, \textit{validity}, and \textit{totality} properties of the asynchronous common subset protocol.

cs.DC↗

arXiv2024

Golden-Retriever: High-Fidelity Agentic Retrieval Augmented Generation for Industrial Knowledge Base

This paper introduces Golden-Retriever, designed to efficiently navigate vast industrial knowledge bases, overcoming challenges in traditional LLM fine-tuning and RAG frameworks with domain-specific jargon and context interpretation. Golden-Retriever incorporates a reflection-based question augmentation step before document retrieval, which involves identifying jargon, clarifying its meaning based on context, and augmenting the question accordingly. Specifically, our method extracts and lists all jargon and abbreviations in the input question, determines the context against a pre-defined list, and queries a jargon dictionary for extended definitions and descriptions. This comprehensive augmentation ensures the RAG framework retrieves the most relevant documents by providing clear context and resolving ambiguities, significantly improving retrieval accuracy. Evaluations using three open-source LLMs on a domain-specific question-answer dataset demonstrate Golden-Retriever's superior performance, providing a robust solution for efficiently integrating and querying industrial knowledge bases.

cs.IR↗

arXiv2024

Go Beyond Black-box Policies: Rethinking the Design of Learning Agent for Interpretable and Verifiable HVAC Control

Recent research has shown the potential of Model-based Reinforcement Learning (MBRL) to enhance energy efficiency of Heating, Ventilation, and Air Conditioning (HVAC) systems. However, existing methods rely on black-box thermal dynamics models and stochastic optimizers, lacking reliability guarantees and posing risks to occupant health. In this work, we overcome the reliability bottleneck by redesigning HVAC controllers using decision trees extracted from existing thermal dynamics models and historical data. Our decision tree-based policies are deterministic, verifiable, interpretable, and more energy-efficient than current MBRL methods. First, we introduce a novel verification criterion for RL agents in HVAC control based on domain knowledge. Second, we develop a policy extraction procedure that produces a verifiable decision tree policy. We found that the high dimensionality of the thermal dynamics model input hinders the efficiency of policy extraction. To tackle the dimensionality challenge, we leverage importance sampling conditioned on historical data distributions, significantly improving policy extraction efficiency. Lastly, we present an offline verification algorithm that guarantees the reliability of a control policy. Extensive experiments show that our method saves 68.4% more energy and increases human comfort gain by 14.8% compared to the state-of-the-art method, in addition to an 1127x reduction in computation overhead. Our code and data are available at https://github.com/ryeii/Veri_HVAC

eess.SY↗

arXiv2024

Reward Bound for Behavioral Guarantee of Model-based Planning Agents

Recent years have seen an emerging interest in the trustworthiness of machine learning-based agents in the wild, especially in robotics, to provide safety assurance for the industry. Obtaining behavioral guarantees for these agents remains an important problem. In this work, we focus on guaranteeing a model-based planning agent reaches a goal state within a specific future time step. We show that there exists a lower bound for the reward at the goal state, such that if the said reward is below that bound, it is impossible to obtain such a guarantee. By extension, we show how to enforce preferences over multiple goals.

cs.AI↗

arXiv2023

Data Race Detection Using Large Language Models

Large language models (LLMs) are demonstrating significant promise as an alternate strategy to facilitate analyses and optimizations of high-performance computing programs, circumventing the need for resource-intensive manual tool creation. In this paper, we explore a novel LLM-based data race detection approach combining prompting engineering and fine-tuning techniques. We create a dedicated dataset named DRB-ML, which is derived from DataRaceBench, with fine-grain labels showing the presence of data race pairs and their associated variables, line numbers, and read/write information. DRB-ML is then used to evaluate representative LLMs and fine-tune open-source ones. Our experiment shows that LLMs can be a viable approach to data race detection. However, they still cannot compete with traditional data race detection tools when we need detailed information about variable pairs causing data races.

cs.LG↗

arXiv2023

HPC-GPT: Integrating Large Language Model for High-Performance Computing

Large Language Models (LLMs), including the LLaMA model, have exhibited their efficacy across various general-domain natural language processing (NLP) tasks. However, their performance in high-performance computing (HPC) domain tasks has been less than optimal due to the specialized expertise required to interpret the model responses. In response to this challenge, we propose HPC-GPT, a novel LLaMA-based model that has been supervised fine-tuning using generated QA (Question-Answer) instances for the HPC domain. To evaluate its effectiveness, we concentrate on two HPC tasks: managing AI models and datasets for HPC, and data race detection. By employing HPC-GPT, we demonstrate comparable performance with existing methods on both tasks, exemplifying its excellence in HPC-related scenarios. Our experiments on open-source benchmarks yield extensive results, underscoring HPC-GPT's potential to bridge the performance gap between LLMs and HPC-specific tasks. With HPC-GPT, we aim to pave the way for LLMs to excel in HPC domains, simplifying the utilization of language models in complex computing applications.

cs.DC↗

arXiv2023

Optimizing Irrigation Efficiency using Deep Reinforcement Learning in the Field

Agricultural irrigation is a significant contributor to freshwater consumption. However, the current irrigation systems used in the field are not efficient. They rely mainly on soil moisture sensors and the experience of growers, but do not account for future soil moisture loss. Predicting soil moisture loss is challenging because it is influenced by numerous factors, including soil texture, weather conditions, and plant characteristics. This paper proposes a solution to improve irrigation efficiency, which is called DRLIC. DRLIC is a sophisticated irrigation system that uses deep reinforcement learning (DRL) to optimize its performance. The system employs a neural network, known as the DRL control agent, which learns an optimal control policy that considers both the current soil moisture measurement and the future soil moisture loss. We introduce an irrigation reward function that enables our control agent to learn from previous experiences. However, there may be instances where the output of our DRL control agent is unsafe, such as irrigating too much or too little water. To avoid damaging the health of the plants, we implement a safety mechanism that employs a soil moisture predictor to estimate the performance of each action. If the predicted outcome is deemed unsafe, we perform a relatively-conservative action instead. To demonstrate the real-world application of our approach, we developed an irrigation system that comprises sprinklers, sensing and control nodes, and a wireless network. We evaluate the performance of DRLIC by deploying it in a testbed consisting of six almond trees. During a 15-day in-field experiment, we compared the water consumption of DRLIC with a widely-used irrigation scheme. Our results indicate that DRLIC outperformed the traditional irrigation method by achieving a water savings of up to 9.52%.

cs.LG↗

arXiv2023

Multi-zone HVAC Control with Model-Based Deep Reinforcement Learning

In this paper, we conduct a set of experiments to analyze the limitations of current MBRL-based HVAC control methods, in terms of model uncertainty and controller effectiveness. Using the lessons learned, we develop MB2C, a novel MBRL-based HVAC control system that can achieve high control performance with excellent sample efficiency. MB2C learns the building dynamics by employing an ensemble of environment-conditioned neural networks. It then applies a new control method, Model Predictive Path Integral (MPPI), for HVAC control. It produces candidate action sequences by using an importance sampling weighted algorithm that scales better to high state and action dimensions of multi-zone buildings. We evaluate MB2C using EnergyPlus simulations in a five-zone office building. The results show that MB2C can achieve 8.23% more energy savings compared to the state-of-the-art MBRL solution while maintaining similar thermal comfort. MB2C can reduce the training data set by an order of magnitude (10.52x) while achieving comparable performance to MFRL approaches.

eess.SY↗

arXiv2023

Exploring Deep Reinforcement Learning for Holistic Smart Building Control

In this paper, we take a holistic approach to deal with the tradeoffs between energy use and comfort in commercial buildings. We developed a system called OCTOPUS, which employs a novel deep reinforcement learning (DRL) framework that uses a data-driven approach to find the optimal control sequences of all building's subsystems, including HVAC, lighting, blind and window systems. The DRL architecture includes a novel reward function that allows the framework to explore the tradeoffs between energy use and users' comfort, while at the same time enabling the solution of the high-dimensional control problem due to the interactions of four different building subsystems. In order to cope with OCTOPUS's data training requirements, we argue that calibrated simulations that match the target building operational points are the vehicle to generate enough data to be able to train our DRL framework to find the control solution for the target building. In our work, we trained OCTOPUS with 10-year weather data and a building model that is implemented in the EnergyPlus building simulator, which was calibrated using data from a real production building. Through extensive simulations, we demonstrate that OCTOPUS can achieve 14.26% and 8.1% energy savings compared with the state-of-the-art rule-based method in a LEED Gold Certified building and the latest DRL-based method available in the literature respectively, while maintaining human comfort within a desired range.

cs.LG↗