arXiv subjects

Zihan Wu

Publications and source records attributed to Zihan Wu.

At least 19 recordsLinked to original sources

arXiv2026

A Unified Issue Resolution Benchmark for Requirement Clarification, Planning, and Code Generation for Coding Agents

Large language model-powered coding agents are increasingly used to modify existing code repositories, for example, by adding features or fixing bugs. Yet existing repository-level benchmarks typically evaluate only whether the final patch passes tests. Satisfying a user request requires a long chain of interdependent reasoning and decisions: an agent must recover explicit and implicit requirements, formulate a repository-grounded implementation plan, and translate it into correct code. A pass/fail outcome cannot characterize how an unsuccessful trajectory diverges from the requirements and implementation process needed for a correct patch. To address this gap, we introduce SWE-RPG, a repository-level benchmark that combines executable patch evaluation with validated ground-truth references (GTs) for (1) Requirement Clarification and (2) Implementation Planning. These intermediate GTs support retrospective, GT-aligned diagnosis of complete coding-agent trajectories across clarification, planning, code generation, and artifact submission. SWE-RPG comprises 163 tasks from 31 Python and Java repositories, including 113 bug fixes and 50 feature additions. We evaluate 3 coding agents, including Claude Code, Codex, and OpenCode, with 6 large language model backends, including Claude-Sonnet-5 and GPT-5.6-Terra. Results show that the evaluated popular coding agents still struggle to implement user requests in existing repositories, achieving an average resolved rate of only 31.5% on SWE-RPG. Intermediate-GT diagnosis further identifies implicit requirement recovery as the main bottleneck, accounting for 24.5%--46.0% of agent runs. This result suggests implicit-requirement recovery as a key candidate direction for improving coding agents. The benchmark data and evaluation code are available at https://github.com/Xin-Zhou-smu/SWE-RPG-Bench.

cs.SE↗

arXiv2026

Exploring the Design Space of LLM-Based Programming Support in CS Education: A Scoping Review through the Lens of Assistance Governance

As large language models (LLMs) become integrated into programming education, learner-facing systems increasingly differ in how that assistance is bounded, enacted, and controlled. These governance decisions are often described implicitly, making it difficult to compare systems in educationally meaningful ways. To address this gap, we conduct a scoping review and qualitative synthesis of 90 peer-reviewed LLM-based programming support systems in CS education. We analyze assistance governance through three dimensions, which we refer to collectively as PEA: Policy, capturing what forms of help are allowed or restricted; Enforcement, capturing how those boundaries are operationalized through interaction and system behavior; and Authority, capturing who can configure, adapt, or override them during use. Our findings show that systems often share similar pedagogical goals, but implement those goals through varied enforcement mechanisms. At the same time, authority remains highly centralized in system logic, with fewer systems giving learners or instructors runtime control. This work contributes PEA as a three-dimensional analytic lens, a governance codebook empirically refined within these dimensions, and a map of underexplored configurations in the current design space of LLM-based programming support. By making these explicit and comparable, PEA offers a vocabulary for analyzing existing systems and designing future tools that are pedagogically bounded, configurable, and accountable.

cs.HC↗

arXiv2026

What Resolve Rate Hides: Trajectory Structure Diagnostics for Coding Agents

Coding agents are ranked almost entirely by resolve rate: whether their final patch passes the target tests. Yet two agents can reach the same outcome through very different processes, and a single pass/fail label says nothing about why a run failed or why an accepted run spent extra steps, time, or tokens. This process evidence lives in the trajectory, which records a run's searches, reads, edits, tool calls, validation, and reversions. However, raw traces are heterogeneous and hard to compare across runs. We present TraceProbe, a trajectory-diagnostic framework that recovers what resolve rate hides. TraceProbe normalizes each raw run into a canonical nine-type action taxonomy with deterministic effect labels, then applies two rule-based modules: Insight names single-trajectory anti-patterns adapted from established debugging practice (e.g., search loops, verification skips), while Converge aligns pairs of runs and classifies where their behavior diverges under controlled references. Applying TraceProbe to 2,500 trajectories from five production settings on SWE-Bench Verified, we find that (i) file choice is too coarse to separate success from failure, whereas function selection and completion behavior localize it; (ii) Insight anti-patterns act mainly as corpus-level difficulty clues, with search loops the most stable; and (iii) even resolved runs differ in how quickly they reach relevant code and how much failed work they incur. Trajectory structure thus adds auditable diagnostic context to outcomes by localizing inspection targets, suggesting failure hypotheses, and prioritizing runs for review.

cs.SE↗

arXiv2026

P$^2$RAG: Efficient Privacy-Preserving RAG Service Supporting Arbitrary Top-$k$ Retrieval

Retrieval-Augmented Generation (RAG) enables large language models to use external knowledge, but outsourcing the RAG service raises privacy concerns for both data owners and users. Privacy-preserving RAG systems address these concerns by performing secure top-$k$ retrieval, which is typically implemented using secure sorting to identify relevant documents. However, existing systems face challenges supporting arbitrary $k$ due to their inability to change $k$, new security issues, and in particular, efficiency degradation with large $k$. This is a significant limitation because applications such as finance, law, and healthcare require a $k$ that is large enough to cause huge overhead for existing systems. Also, modern long-context models generally achieve higher accuracy with larger retrieval sets. We propose P$^2$RAG, an efficient privacy-preserving RAG service that supports arbitrary top-$k$ retrieval. Unlike existing systems, P$^2$RAG avoids sorting candidate documents. Instead, it uses an interactive bisection method to determine the set of top-$k$ documents. For security, P$^2$RAG uses secret sharing on two semi-honest non-colluding servers to protect the data owner's database and the user's prompt. It enforces restrictions and verification to defend against malicious users and tightly bounds the information leakage of the database. The experiments show that P$^2$RAG is 3--300$\times$ faster than the state-of-the-art PRAG for $k = 16$--$1024$.

cs.CR↗

arXiv2026

Exploring the Role of Tracing in AI-Supported Planning for Algorithmic Reasoning

AI-powered planning tools show promise in supporting programming learners by enabling early, formative feedback on their thinking processes prior to coding. To date, however, most AI-supported planning tools rely on students' natural-language explanations, using LLMs to interpret learners' descriptions of their algorithmic intent. Prior to the emergence of LLM-based systems, CS education research extensively studied trace-based planning in pen-and-paper settings, demonstrating that reasoning through stepwise execution with explicit state transitions helps learners build and refine mental models of program behavior. Despite its potential, little is known about how tracing interacts with AI-mediated feedback and whether integrating tracing into AI-supported planning tools leads to different learning processes or interaction dynamics compared to natural-language-based planning alone. We study how requiring learners to produce explicit execution traces with an AI-supported planning tool affects their algorithmic reasoning. In a between-subjects study with 20 students, tracing shifted learners away from code-like, line-by-line descriptions toward more goal-driven reasoning about program behavior. Moreover, it led to more consistent partially correct solutions, although final coding performance remained comparable across conditions. Notably, tracing did not significantly affect the quality or reliability of LLM-generated feedback. These findings reveal tradeoffs in combining tracing with AI-supported planning and inform design guidelines for integrating natural language, tracing, and coding to support iterative reasoning throughout the programming process.

cs.HC↗

arXiv2026

MulVul: Retrieval-augmented Multi-Agent Code Vulnerability Detection via Cross-Model Prompt Evolution

Large Language Models (LLMs) struggle to automate real-world vulnerability detection due to two key limitations: the heterogeneity of vulnerability patterns undermines the effectiveness of a single unified model, and manual prompt engineering for massive weakness categories is unscalable. To address these challenges, we propose \textbf{MulVul}, a retrieval-augmented multi-agent framework designed for precise and broad-coverage vulnerability detection. MulVul adopts a coarse-to-fine strategy: a \emph{Router} agent first predicts the top-$k$ coarse categories and then forwards the input to specialized \emph{Detector} agents, which identify the exact vulnerability types. Both agents are equipped with retrieval tools to actively source evidence from vulnerability knowledge bases to mitigate hallucinations. Crucially, to automate the generation of specialized prompts, we design \emph{Cross-Model Prompt Evolution}, a prompt optimization mechanism where a generator LLM iteratively refines candidate prompts while a distinct executor LLM validates their effectiveness. This decoupling mitigates the self-correction bias inherent in single-model optimization. Evaluated on 130 CWE types, MulVul achieves 34.79\% Macro-F1, outperforming the best baseline by 41.5\%. Ablation studies validate cross-model prompt evolution, which boosts performance by 51.6\% over manual prompts by effectively handling diverse vulnerability patterns.

cs.SE↗

arXiv2025

ExpStar: Towards Automatic Commentary Generation for Multi-discipline Scientific Experiments

Experiment commentary is crucial in describing the experimental procedures, delving into underlying scientific principles, and incorporating content-related safety guidelines. In practice, human teachers rely heavily on subject-specific expertise and invest significant time preparing such commentary. To address this challenge, we introduce the task of automatic commentary generation across multi-discipline scientific experiments. While recent progress in large multimodal models (LMMs) has demonstrated promising capabilities in video understanding and reasoning, their ability to generate fine-grained and insightful experiment commentary remains largely underexplored. In this paper, we make the following contributions: (i) We construct \textit{ExpInstruct}, the first dataset tailored for experiment commentary generation, featuring over 7\textit{K} step-level commentaries across 21 scientific subjects from 3 core disciplines (\ie, science, healthcare and engineering). Each sample includes procedural descriptions along with potential scientific principles (\eg, chemical equations and physical laws) and safety guidelines. (ii) We propose ExpStar, an automatic experiment commentary generation model that leverages a retrieval-augmented mechanism to adaptively access, evaluate, and utilize external knowledge. (iii) Extensive experiments show that our ExpStar substantially outperforms 14 leading LMMs, which highlights the superiority of our dataset and model. We believe that ExpStar holds great potential for advancing AI-assisted scientific experiment instruction.

cs.CV↗

arXiv2025

Scalable Co-Clustering for Large-Scale Data through Dynamic Partitioning and Hierarchical Merging

Co-clustering simultaneously clusters rows and columns, revealing more fine-grained groups. However, existing co-clustering methods suffer from poor scalability and cannot handle large-scale data. This paper presents a novel and scalable co-clustering method designed to uncover intricate patterns in high-dimensional, large-scale datasets. Specifically, we first propose a large matrix partitioning algorithm that partitions a large matrix into smaller submatrices, enabling parallel co-clustering. This method employs a probabilistic model to optimize the configuration of submatrices, balancing the computational efficiency and depth of analysis. Additionally, we propose a hierarchical co-cluster merging algorithm that efficiently identifies and merges co-clusters from these submatrices, enhancing the robustness and reliability of the process. Extensive evaluations validate the effectiveness and efficiency of our method. Experimental results demonstrate a significant reduction in computation time, with an approximate 83% decrease for dense matrices and up to 30% for sparse matrices.

cs.DC↗

arXiv2025

Learner and Instructor Needs in AI-Supported Programming Learning Tools: Design Implications for Features and Adaptive Control

AI-supported tools can help learners overcome challenges in programming education by providing adaptive assistance. However, existing research often focuses on individual tools rather than deriving broader design recommendations. A key challenge in designing these systems is balancing learner control with system-driven guidance. To explore user preferences for AI-supported programming learning tools, we conducted a participatory design study with 15 undergraduate novice programmers and 10 instructors to gather insights on their desired help features and control preferences, as well as a follow-up survey with 172 introductory programming students. Our qualitative findings show that learners prefer help that is encouraging, incorporates visual aids, and includes peer-related insights, whereas instructors prioritize scaffolding that reflects learners' progress and reinforces best practices. Both groups favor shared control, though learners generally prefer more autonomy, while instructors lean toward greater system guidance to prevent cognitive overload. Additionally, our interviews revealed individual differences in control preferences. Based on our findings, we propose design guidelines for AI-supported programming tools, particularly regarding user-centered help features and adaptive control mechanisms. Our work contributes to the human-centered design of AI-supported learning environments by informing the development of systems that effectively balance autonomy and guidance, enhancing AI-supported educational tools for programming and beyond.

cs.HC↗

arXiv2025

Dynamic Influence Tracker: Measuring Time-Varying Sample Influence During Training

Existing methods for measuring training sample influence on models only provide static, overall measurements, overlooking how sample influence changes during training. We propose Dynamic Influence Tracker (DIT), which captures the time-varying sample influence across arbitrary time windows during training. DIT offers three key insights: 1) Samples show different time-varying influence patterns, with some samples important in the early training stage while others become important later. 2) Sample influences show a weak correlation between early and late stages, demonstrating that the model undergoes distinct learning phases with shifting priorities. 3) Analyzing influence during the convergence period provides more efficient and accurate detection of corrupted samples than full-training analysis. Supported by theoretical guarantees without assuming loss convexity or model convergence, DIT significantly outperforms existing methods, achieving up to 0.99 correlation with ground truth and above 98\% accuracy in detecting corrupted samples in complex architectures.

stat.ML↗

arXiv2025

LiveVal: Time-aware Data Valuation via Adaptive Reference Points

Time-aware data valuation enhances training efficiency and model robustness, as early detection of harmful samples could prevent months of wasted computation. However, existing methods rely on model retraining or convergence assumptions or fail to capture long-term training dynamics. We propose LiveVal, an efficient time-aware data valuation method with three key designs: 1) seamless integration with SGD training for efficient data contribution monitoring; 2) reference-based valuation with normalization for reliable benchmark establishment; and 3) adaptive reference point selection for real-time updating with optimized memory usage. We establish theoretical guarantees for LiveVal's stability and prove that its valuations are bounded and directionally aligned with optimization progress. Extensive experiments demonstrate that LiveVal provides efficient data valuation across different modalities and model scales, achieving 180 speedup over traditional methods while maintaining robust detection performance.

cs.LG↗

arXiv2025

Classic4Children: Adapting Chinese Literary Classics for Children with Large Language Model

Chinese literary classics hold significant cultural and educational value, offering deep insights into morality, history, and human nature. These works often include classical Chinese and complex narratives, making them difficult for children to read. To bridge this gap, we introduce a child-friendly literary adaptation (CLA) task to adapt the Chinese literary classic into engaging and accessible text for children. However, recent large language models (LLMs) overlook children's reading preferences (\ie, vivid character portrayals, concise narrative structures, and appropriate readability), which poses challenges in CLA. In this paper, we propose a method called InstructChild, which augments the LLM with these preferences for adaptation. Specifically, we first obtain the characters' personalities and narrative structure as additional information for fine-grained instruction tuning. Then, we devise a readability metric as the reward to align the LLM with the children's reading level. Finally, a lookahead decoding strategy is applied to improve the readability of the generated text during inference. To support the evaluation of CLA task, we construct the Classic4Children dataset, which comprises both the original and child-friendly versions of the Four Great Classical Novels of Chinese literature. Experimental results show that our InstructChild significantly improves automatic and human evaluation performance.

cs.CL↗

arXiv2025

Personalized Parsons Puzzles as Scaffolding Enhance Practice Engagement Over Just Showing LLM-Powered Solutions

As generative AI products could generate code and assist students with programming learning seamlessly, integrating AI into programming education contexts has driven much attention. However, one emerging concern is that students might get answers without learning from the LLM-generated content. In this work, we deployed the LLM-powered personalized Parsons puzzles as scaffolding to write-code practice in a Python learning classroom (PC condition) and conducted an 80-minute randomized between-subjects study. Both conditions received the same practice problems. The only difference was that when requesting help, the control condition showed students a complete solution (CC condition), simulating the most traditional LLM output. Results indicated that students who received personalized Parsons puzzles as scaffolding engaged in practicing significantly longer than those who received complete solutions when struggling.

cs.HC↗

arXiv2024

Insights from Social Shaping Theory: The Appropriation of Large Language Models in an Undergraduate Programming Course

The capability of large language models (LLMs) to generate, debug, and explain code has sparked the interest of researchers and educators in undergraduate programming, with many anticipating their transformative potential in programming education. However, decisions about why and how to use LLMs in programming education may involve more than just the assessment of an LLM's technical capabilities. Using the social shaping of technology theory as a guiding framework, our study explores how students' social perceptions influence their own LLM usage. We then examine the correlation of self-reported LLM usage with students' self-efficacy and midterm performances in an undergraduate programming course. Triangulating data from an anonymous end-of-course student survey (n = 158), a mid-course self-efficacy survey (n=158), student interviews (n = 10), self-reported LLM usage on homework, and midterm performances, we discovered that students' use of LLMs was associated with their expectations for their future careers and their perceptions of peer usage. Additionally, early self-reported LLM usage in our context correlated with lower self-efficacy and lower midterm scores, while students' perceived over-reliance on LLMs, rather than their usage itself, correlated with decreased self-efficacy later in the course.

cs.HC↗

arXiv2024

CodeTailor: LLM-Powered Personalized Parsons Puzzles for Engaging Support While Learning Programming

Learning to program can be challenging, and providing high-quality and timely support at scale is hard. Generative AI and its products, like ChatGPT, can create a solution for most intro-level programming problems. However, students might use these tools to just generate code for them, resulting in reduced engagement and limited learning. In this paper, we present CodeTailor, a system that leverages a large language model (LLM) to provide personalized help to students while still encouraging cognitive engagement. CodeTailor provides a personalized Parsons puzzle to support struggling students. In a Parsons puzzle, students place mixed-up code blocks in the correct order to solve a problem. A technical evaluation with previous incorrect student code snippets demonstrated that CodeTailor could deliver high-quality (correct, personalized, and concise) Parsons puzzles based on their incorrect code. We conducted a within-subjects study with 18 novice programmers. Participants perceived CodeTailor as more engaging than just receiving an LLM-generated solution (the baseline condition). In addition, participants applied more supported elements from the scaffolded practice to the posttest when using CodeTailor than baseline. Overall, most participants preferred using CodeTailor versus just receiving the LLM-generated code for learning. Qualitative observations and interviews also provided evidence for the benefits of CodeTailor, including thinking more about solution construction, fostering continuity in learning, promoting reflection, and boosting confidence. We suggest future design ideas to facilitate active learning opportunities with generative AI techniques.

cs.CY↗

arXiv2024

Evaluating Micro Parsons Problems as Exam Questions

Parsons problems are a type of programming activity that present learners with blocks of existing code and requiring them to arrange those blocks to form a program rather than write the code from scratch. Micro Parsons problems extend this concept by having students assemble segments of code to form a single line of code rather than an entire program. Recent investigations into micro Parsons problems have primarily focused on supporting learners leaving open the question of micro Parsons efficacy as an exam item and how students perceive it when preparing for exams. To fill this gap, we included a variety of micro Parsons problems on four exams in an introductory programming course taught in Python. We use Item Response Theory to investigate the difficulty of the micro Parsons problems as well as the ability of the questions to differentiate between high and low ability students. We then compare these results to results for related questions where students are asked to write a single line of code from scratch. Finally, we conduct a thematic analysis of the survey responses to investigate how students' perceptions of micro Parsons both when practicing for exams and as they appear on exams.

cs.HC↗

arXiv2024

Exploring the Distinctiveness and Fidelity of the Descriptions Generated by Large Vision-Language Models

Large Vision-Language Models (LVLMs) are gaining traction for their remarkable ability to process and integrate visual and textual data. Despite their popularity, the capacity of LVLMs to generate precise, fine-grained textual descriptions has not been fully explored. This study addresses this gap by focusing on \textit{distinctiveness} and \textit{fidelity}, assessing how models like Open-Flamingo, IDEFICS, and MiniGPT-4 can distinguish between similar objects and accurately describe visual features. We proposed the Textual Retrieval-Augmented Classification (TRAC) framework, which, by leveraging its generative capabilities, allows us to delve deeper into analyzing fine-grained visual description generation. This research provides valuable insights into the generation quality of LVLMs, enhancing the understanding of multimodal language models. Notably, MiniGPT-4 stands out for its better ability to generate fine-grained descriptions, outperforming the other two models in this aspect. The code is provided at \url{https://anonymous.4open.science/r/Explore_FGVDs-E277}.

cs.CV↗

arXiv2024

LMEraser: Large Model Unlearning through Adaptive Prompt Tuning

To address the growing demand for privacy protection in machine learning, we propose a novel and efficient machine unlearning approach for \textbf{L}arge \textbf{M}odels, called \textbf{LM}Eraser. Existing unlearning research suffers from entangled training data and complex model architectures, incurring extremely high computational costs for large models. LMEraser takes a divide-and-conquer strategy with a prompt tuning architecture to isolate data influence. The training dataset is partitioned into public and private datasets. Public data are used to train the backbone of the model. Private data are adaptively clustered based on their diversity, and each cluster is used to optimize a prompt separately. This adaptive prompt tuning mechanism reduces unlearning costs and maintains model performance. Experiments demonstrate that LMEraser achieves a $100$-fold reduction in unlearning costs without compromising accuracy compared to prior work. Our code is available at: \url{https://github.com/lmeraser/lmeraser}.

cs.LG↗