arXiv subjects

Xiufeng Xu

Publications and source records attributed to Xiufeng Xu.

6 recordsLinked to original sources

arXiv2026

Neurosymbolic Repo-level Code Localization

Code localization is a cornerstone of autonomous software engineering. Recent advancements have achieved impressive performance on real-world issue benchmarks. However, we identify a critical yet overlooked bias: these benchmarks are saturated with keyword references (e.g. file paths, function names), encouraging models to rely on superficial lexical matching rather than genuine structural reasoning. We term this phenomenon the Keyword Shortcut. To address this, we formalize the challenge of Keyword-Agnostic Logical Code Localization (KA-LCL) and introduce KA-LogicQuery, a diagnostic benchmark requiring structural reasoning without any naming hints. Our evaluation reveals a catastrophic performance drop of state-of-the-art approaches on KA-LogicQuery, exposing their lack of deterministic reasoning capabilities. We propose LogicLoc, a novel agentic framework that combines large language models with the rigorous logical reasoning of Datalog for precise localization. LogicLoc extracts program facts from the codebase and leverages an LLM to synthesize Datalog programs, with parser-gated validation and mutation-based intermediate-rule diagnostic feedback to ensure correctness and efficiency. The validated programs are executed by a high-performance inference engine, enabling accurate and verifiable localization in a fully automated, closed-loop workflow. Experimental results demonstrate that LogicLoc significantly outperforms SOTA methods on KA-LogicQuery while maintaining competitive performance on popular issue-driven benchmarks. Notably, LogicLoc attains superior performance with significantly lower token consumption and faster execution by offloading structural traversal to a deterministic engine, reducing the overhead of iterative LLM inference.

cs.SE↗

arXiv2024

Detecting Multi-Parameter Constraint Inconsistencies in Python Data Science Libraries

Modern AI- and Data-intensive software systems rely heavily on data science and machine learning libraries that provide essential algorithmic implementations and computational frameworks. These libraries expose complex APIs whose correct usage has to follow constraints among multiple interdependent parameters. Developers using these APIs are expected to learn about the constraints through the provided documentations and any discrepancy may lead to unexpected behaviors. However, maintaining correct and consistent multi-parameter constraints in API documentations remains a significant challenge for API compatibility and reliability. To address this challenge, we propose MPDetector, for detecting inconsistencies between code and documentation, specifically focusing on multi-parameter constraints. MPDetector identifies these constraints at the code level by exploring execution paths through symbolic execution and further extracts corresponding constraints from documentation using large language models (LLMs). We propose a customized fuzzy constraint logic to reconcile the unpredictability of LLM outputs and detects logical inconsistencies between the code and documentation constraints. We collected and constructed two datasets from four popular data science libraries and evaluated MPDetector on them. The results demonstrate that MPDetector can effectively detect inconsistency issues with the precision of 92.8%. We further reported 14 detected inconsistency issues to the library developers, who have confirmed 11 issues at the time of writing.

cs.SE↗

arXiv2024

Stochastic behavior of an n-node blockchain under cyber attacks from multiple hackers with random re-setting times

This paper investigates the stochastic behavior of an n-node blockchain which is continuously monitored and faces non-stop cyber attacks from multiple hackers. The blockchain will start being re-set once hacking is detected, forfeiting previous efforts of all hackers. It is assumed the re-setting process takes a random amount of time. Multiple independent hackers will keep attempting to hack into the blockchain until one of them succeeds. For arbitrary distributions of the hacking times, detecting times, and re-setting times, we derive the instantaneous functional probability, the limiting functional probability, and the mean functional time of the blockchain. Moreover, we establish that these quantities are increasing functions of the number of nodes, formalizing the intuition that the more nodes a blockchain has the more secure it is.

stat.AP↗

arXiv2024

LLMs for Relational Reasoning: How Far are We?

Large language models (LLMs) have revolutionized many areas (e.g. natural language processing, software engineering, etc.) by achieving state-of-the-art performance on extensive downstream tasks. Aiming to achieve robust and general artificial intelligence, there has been a surge of interest in investigating the reasoning ability of the LLMs. Whereas the textual and numerical reasoning benchmarks adopted by previous works are rather shallow and simple, it is hard to conclude that the LLMs possess strong reasoning ability by merely achieving positive results on these benchmarks. Recent efforts have demonstrated that the LLMs are poor at solving sequential decision-making problems that require common-sense planning by evaluating their performance on the reinforcement learning benchmarks. In this work, we conduct an in-depth assessment of several state-of-the-art LLMs' reasoning ability based on the inductive logic programming (ILP) benchmark, which is broadly recognized as a representative and challenging measurement for evaluating logic program induction/synthesis systems as it requires inducing strict cause-effect logic to achieve robust deduction on independent and identically distributed (IID) and out-of-distribution (OOD) test samples. Our evaluations illustrate that compared with the neural program induction systems which are much smaller in model size, the state-of-the-art LLMs are much poorer in terms of reasoning ability by achieving much lower performance and generalization using either natural language prompting or truth-value matrix prompting.

cs.AI↗

arXiv2023

CompSuite: A Dataset of Java Library Upgrade Incompatibility Issues

Modern software systems heavily rely on external libraries developed by third-parties to ensure efficient development. However, frequent library upgrades can lead to compatibility issues between the libraries and their client systems. In this paper, we introduce CompSuite, a dataset that includes 123 real-world Java client-library pairs where upgrading the library causes an incompatibility issue in the corresponding client. Each incompatibility issue in CompSuite is associated with a test case authored by the developers, which can be used to reproduce the issue. The dataset also provides a command-line interface that simplifies the execution and validation of each issue. With this infrastructure, users can perform an inspection of any incompatibility issue with the push of a button, or reproduce an issue step-by-step for a more detailed investigation. We make CompSuite publicly available to promote open science. We believe that various software analysis techniques, such as compatibility checking, debugging, and regression test selection, can benefit from CompSuite.

cs.SE↗

arXiv2022

Instantaneous and limiting behavior of an n-node blockchain under cyber attacks from a single hacker

We investigate the instantaneous and limiting behavior of an n-node blockchain which is under continuous monitoring of the IT department of a company but faces non-stop cyber attacks from a single hacker. The blockchain is functional as far as no data stored on it has been changed, deleted, or locked. Once the IT department detects the attack from the hacker, it will immediately re-set the blockchain, rendering all previous efforts of the hacker in vain. The hacker will not stop until the blockchain is dysfunctional. For arbitrary distributions of the hacking times and detecting times, we derive the limiting functional probability, instantaneous functional probability, and mean functional time of the blockchain. We also show that all these quantities are increasing functions of the number of nodes, substantiating the intuition that the more nodes a blockchain has, the harder it is for a hacker to succeed in a cyber attack.

cs.CR↗