Search

Large Language Models (LLMs) have become a focal point of research across various domains, including software engineering, where their capabilities are increasingly leveraged. Recent studies have explored the integration of LLMs into software development tools and frameworks, revealing their potential to enhance performance in text and code-related tasks. Log level is a key part of a logging statement that allows software developers control the information recorded during system runtime. Given that log messages often mix natural language with code-like variables, LLMs' language translation abilities could be applied to determine the suitable verbosity level for logging statements. In this paper, we undertake a detailed empirical analysis to investigate the impact of characteristics and learning paradigms on the performance of 12 open-source LLMs in log level suggestion. We opted for open-source models because they enable us to utilize in-house code while effectively protecting sensitive information and maintaining data security. We examine several prompting strategies, including Zero-shot, Few-shot, and fine-tuning techniques, across different LLMs to identify the most effective combinations for accurate log level suggestions. Our research is supported by experiments conducted on 9 large-scale Java systems. The results indicate that although smaller LLMs can perform effectively with appropriate instruction and suitable techniques, there is still considerable potential for improvement in their ability to suggest log levels.

The code review comment (CRC) is pivotal in the process of modern code review. It provides reviewers with the opportunity to identify potential bugs, offer constructive feedback, and suggest improvements. Clear and concise code review comments (CRCs) facilitate the communication between developers and is crucial to the correct understanding of the issues identified and proposed solutions. Despite the importance of CRCs' clarity, there is still a lack of guidelines on what constitutes a good clarity and how to evaluate it. In this paper, we conduct a comprehensive study on understanding and evaluating the clarity of CRCs. We first derive a set of attributes related to the clarity of CRCs, namely RIE attributes (i.e., Relevance, Informativeness, and Expression), as well as their corresponding evaluation criteria based on our literature review and survey with practitioners. We then investigate the clarity of CRCs in open-source projects written in nine programming languages and find that a large portion (i.e., 28.8%) of the CRCs lack the clarity in at least one of the attributes. Finally, we propose ClearCRC, an automated framework that evaluates the clarity of CRCs. Experimental results show that ClearCRC can effectively evaluate the clarity of CRCs and outperform the baselines.

We consider a frictionless system coupled to an external Markovian environment. The quantum and classical evolution of such systems are described by the Lindblad and the Fokker-Planck equation respectively. We show that when such a system is given by an at most quadratically growing Hamiltonian and at most linearly growing real jump functions, the quantum and classical evolutions remain close on time scales much longer than Ehrenfest time. In particular, we show that the evolution of a density matrix by the Lindblad equation is close in trace norm to the quantization of the corresponding evolution by the Fokker-Planck equation. Such agreement improves upon recent results [arXiv:2403.09345, arXiv:2306.13717, arXiv:2307.05326], which proved long time agreement in weaker norms., Comment: 29 pages, 2 figures

Though many approaches have been proposed for Automated Program Repair (APR) and indeed achieved remarkable performance, they still have limitations in fixing bugs that require analyzing and reasoning about the logic of the buggy program. Recently, large language models (LLMs) instructed by prompt engineering have attracted much attention for their powerful ability to address many kinds of tasks including bug-fixing. However, the quality of the prompt will highly affect the ability of LLMs and manually constructing high-quality prompts is a costly endeavor. To address this limitation, we propose a self-directed LLM-based automated program repair, ThinkRepair, with two main phases: collection phase and fixing phase. The former phase automatically collects various chains of thoughts that constitute pre-fixed knowledge by instructing LLMs with the Chain-of-Thought (CoT) prompt. The latter phase targets fixing a bug by first selecting examples for few-shot learning and second automatically interacting with LLMs, optionally appending with feedback of testing information. Evaluations on two widely studied datasets (Defects4J and QuixBugs) by comparing ThinkRepair with 12 SOTA APRs indicate the priority of ThinkRepair in fixing bugs. Notably, ThinkRepair fixes 98 bugs and improves baselines by 27%-344.4% on Defects4J V1.2. On Defects4J V2.0, ThinkRepair fixes 12-65 more bugs than the SOTA APRs. Additionally, ThinkRepair also makes a considerable improvement on QuixBugs (31 for Java and 21 for Python at most)., Comment: Accepted By ISSTA'24

Pretrained language models have significantly advanced performance across various natural language processing tasks. However, adversarial attacks continue to pose a critical challenge to system built using these models, as they can be exploited with carefully crafted adversarial texts. Inspired by the ability of diffusion models to predict and reduce noise in computer vision, we propose a novel and flexible adversarial defense method for language classification tasks, DiffuseDef, which incorporates a diffusion layer as a denoiser between the encoder and the classifier. During inference, the adversarial hidden state is first combined with sampled noise, then denoised iteratively and finally ensembled to produce a robust text representation. By integrating adversarial training, denoising, and ensembling techniques, we show that DiffuseDef improves over different existing adversarial defense methods and achieves state-of-the-art performance against common adversarial attacks.

Large language models (LLMs) achieve promising results in code generation based on a given natural language description. They have been integrated into open-source projects and commercial products to facilitate daily coding activities. The natural language description in the prompt is crucial for LLMs to comprehend users' requirements. Prior studies uncover that LLMs are sensitive to the changes in the prompts, including slight changes that look inconspicuous. However, the natural language descriptions often vary in real-world scenarios (e.g., different formats, grammar, and wording). Prior studies on the robustness of LLMs are often based on random perturbations and such perturbations may not actually happen. In this paper, we conduct a comprehensive study to investigate how are code LLMs robust to variations of natural language description in real-world scenarios. We summarize 18 categories of perturbations of natural language and 3 combinations of co-occurred categories based on our literature review and an online survey with practitioners. We propose an automated framework, NLPerturbator, which can perform perturbations of each category given a set of prompts. Through a series of experiments on code generation using six code LLMs, we find that the perturbed prompts can decrease the performance of code generation by a considerable margin (e.g., up to 21.2%, and 4.8% to 6.1% on average). Our study highlights the importance of enhancing the robustness of LLMs to real-world variations in the prompts, as well as the essentiality of attentively constructing the prompts.

Software vulnerabilities pose significant risks to the security and integrity of software systems. Prior studies have proposed various approaches to vulnerability detection using deep learning or pre-trained models. However, there is still a lack of detailed explanations for understanding vulnerabilities beyond merely detecting their occurrence, which fails to truly help software developers understand and remediate the issues. Recently, large language models (LLMs) have demonstrated remarkable capabilities in comprehending complex contexts and generating content, presenting new opportunities for both detecting and explaining software vulnerabilities. In this paper, we conduct a comprehensive study to investigate the capabilities of LLMs in both detecting and explaining vulnerabilities, and we propose LLMVulExp, a framework that utilizes LLMs for these tasks. Under specialized fine-tuning for vulnerability explanation, our LLMVulExp not only detects the types of vulnerabilities in the code but also analyzes the code context to generate the cause, location, and repair suggestions for these vulnerabilities. These detailed explanations are crucial for helping developers quickly analyze and locate vulnerability issues, providing essential guidance and reference for effective remediation. We find that LLMVulExp can effectively enable the LLMs to perform vulnerability detection (e.g., achieving over a 90\% F1 score on the SeVC dataset) and provide detailed explanations. We also explore the potential of using advanced strategies such as Chain-of-Thought (CoT) to guide the LLMs in concentrating on vulnerability-prone code, achieving promising results.

Over the past decade, modern code review (MCR) has been deemed as a crucial practice of software quality assurance, which is applied to improve software quality and transfer development knowledge within a software team. Despite its importance, MCR is often a complicated and time-consuming activity for practitioners. In recent years, many studies that are dedicated to the comprehension and the improvement of MCR have been explored so that the MCR activity can be carried out more conveniently and efficiently. To provide researchers and practitioners a clear understanding of the current research status on MCR, this paper conducts a systematic literature review of the past years. Given the collected 231 surveyed papers, this paper makes the following five contributions: First, we analyze the research trends of related MCR studies. Second, we provide a taxonomy for the current MCR, encompassing both Improvement Techniques and Understanding Studies. Third, we present the concrete research progress of each novel MCR methodology and prototype tool. Fourth, we exploit the main empirical insights from empirical study and user study that are helpful to improve MCR. Finally, we sum up unsolved challenges and outline several possible research opportunities in the future., Comment: 62 pages

Large language models for code (i.e., code LLMs) have shown strong code understanding and generation capabilities. To evaluate the capabilities of code LLMs in various aspects, many benchmarks have been proposed (e.g., HumanEval and ClassEval). Code reasoning is one of the most essential abilities of code LLMs, but existing benchmarks for code reasoning are not sufficient. Typically, they focus on predicting the input and output of a program, ignoring the evaluation of the intermediate behavior during program execution, as well as the logical consistency (e.g., the model should not give the correct output if the prediction of execution path is wrong) when performing the reasoning. To address these problems, in this paper, we propose a framework, namely REval, for evaluating code reasoning abilities and consistency of code LLMs with program execution. We utilize existing code benchmarks and adapt them to new benchmarks within our framework. A large-scale empirical study is conducted and most LLMs show unsatisfactory performance on both Runtime Behavior Reasoning (i.e., an average accuracy of 44.4%) and Incremental Consistency Evaluation (i.e., an average IC score of 10.3). Evaluation results of current code LLMs reflect the urgent need for the community to strengthen the code reasoning capability of code LLMs. Our code, data, and \newname leaderboard are available at https://r-eval.github.io., Comment: Accepted by 47th IEEE/ACM International Conference on Software Engineering (ICSE 2025). Our REval leaderboard is available at https://r-eval.github.io

We give a precise microlocal description of the singular profile that forms in the long-time propagation of internal waves in an effectively two-dimensional aquarium. We allow domains with corners, such as polygons appearing in the experimental set ups of Maas et al. This extends the previous work of Dyatlov--Wang--Zworski arXiv:2112.10191, which considered domains with smooth boundary. We show that in addition to singularities that correspond to attractors in the underlying classical dynamics, milder singularities propagate out of the corners as well., Comment: Minor edits and simplifications, 103 pages, 12 figures

Despite the recent remarkable achievement in gaze estimation, efficient and accurate personalization of gaze estimation without labels is a practical problem but rarely touched on in the literature. To achieve efficient personalization, we take inspiration from the recent advances in Natural Language Processing (NLP) by updating a negligible number of parameters, "prompts", at the test time. Specifically, the prompt is additionally attached without perturbing original network and can contain less than 1% of a ResNet-18's parameters. Our experiments show high efficiency of the prompt tuning approach. The proposed one can be 10 times faster in terms of adaptation speed than the methods compared. However, it is non-trivial to update the prompt for personalized gaze estimation without labels. At the test time, it is essential to ensure that the minimizing of particular unsupervised loss leads to the goals of minimizing gaze estimation error. To address this difficulty, we propose to meta-learn the prompt to ensure that its updates align with the goal. Our experiments show that the meta-learned prompt can be effectively adapted even with a simple symmetry loss. In addition, we experiment on four cross-dataset validations to show the remarkable advantages of the proposed method. Code is available at https://github.com/hmarkamcan/TPGaze., Comment: Accepted by AAAI 2024

Developers write logging statements to generate logs that provide run-time information for various tasks. The readability of log messages in the logging statements (i.e., the descriptive text) is rather crucial to the value of the generated logs. Immature log messages may slow down or even obstruct the process of log analysis. Despite the importance of log messages, there is still a lack of standards on what constitutes good readability in log messages and how to write them. In this paper, we conduct a series of interviews with 17 industrial practitioners to investigate their expectations on the readability of log messages. Through the interviews, we derive three aspects related to the readability of log messages, including Structure, Information, and Wording, along with several specific practices to improve each aspect. We validate our findings through a series of online questionnaire surveys and receive positive feedback from the participants. We then manually investigate the readability of log messages in large-scale open source systems and find that a large portion (38.1%) of the log messages have inadequate readability. Motivated by such observation, we further explore the potential of automatically classifying the readability of log messages using deep learning and machine learning models. We find that both deep learning and machine learning models can effectively classify the readability of log messages with a balanced accuracy above 80.0% on average. Our study provides comprehensive guidelines for composing log messages to further improve practitioners' logging practices., Comment: Accepted as a research paper at the 38th IEEE/ACM International Conference on Automated Software Engineering (ASE 2023)

We study a model of internal waves in an effectively 2D aquarium under periodic forcing. In the case when the underlying classical dynamics has sufficiently irrational rotation number, we prove that the energy of the internal waves remains bounded. This involves studying the spectrum of a related 0-th order pseudodifferential operator at spectral parameters corresponding to such dynamics. For the special cases of rectangular and elliptic domains, we give an explicit spectral description of that operator., Comment: 19 pages, 4 figures

Due to the sheer size of software logs, developers rely on automated techniques for log analysis. One of the first and most important steps of automated log analysis is log abstraction, which parses the raw logs into a structured format. Prior log abstraction techniques aim to identify and abstract all the dynamic variables in logs and output a static log template for automated log analysis. However, these abstracted dynamic variables may also contain important information that is useful to different tasks in log analysis. In this paper, we investigate the characteristics of dynamic variables and their importance in practice, and explore the potential of a variable-aware log abstraction technique. Through manual investigations and surveys with practitioners, we find that different categories of dynamic variables record various information that can be important depending on the given tasks, the distinction of dynamic variables in log abstraction can further assist in log analysis. We then propose a deep learning based log abstraction approach, named VALB, which can identify different categories of dynamic variables and preserve the value of specified categories of dynamic variables along with the log templates (i.e., variable-aware log abstraction). Through the evaluation on a widely used log abstraction benchmark, we find that VALB outperforms other state-of-the-art log abstraction techniques on general log abstraction (i.e., when abstracting all the dynamic variables) and also achieves a high variable-aware log abstraction accuracy that further identifies the category of the dynamic variables. Our study highlights the potential of leveraging the important information recorded in the dynamic variables to further improve the process of log analysis.

This paper focuses on Code Generation task that aims at generating relevant code fragments according to given natural language descriptions. In the process of software development, developers often encounter two scenarios. One is requested to write a large amount of repetitive and low-technical code for implementing common functionalities. The other is writing code that depends on specific task requirements, which may necessitate the use of external resources such as documentation or other tools. Therefore, code generation has received a lot of attention among academia and industry for assisting developers in coding. In fact, it has also been one of the key concerns in the field of software engineering to make machines understand users' requirements and write programs on their own. The recent development of deep learning techniques especially pre-training models make the code generation task achieve promising performance. In this paper, we systematically review the current work on deep learning-based code generation and classify the current deep learning-based code generation methods into three categories: methods based on code features, methods incorporated with retrieval, and methods incorporated with post-processing. The first category refers to the methods that use deep learning algorithms for code generation based on code features, and the second and third categories of methods improve the performance of the methods in the first category. In this paper, the existing research results of each category of methods are systematically reviewed, summarized and commented. Besides, the paper summarizes and analyzes the corpus and the popular evaluation metrics used in the existing code generation work. Finally, the paper summarizes the overall literature review and provides a prospect on future research directions worthy of attention., Comment: in Chinese language

We study a model of internal waves under periodic forcing in an effectively 2-dimensional aquarium. When the underlying classical dynamics has sufficiently irrational rotation number, we prove that the solution to the internal waves equation remains bounded in energy space in time. This result is in contrast with the works by Colin de Verdi\`ere--Saint-Raymond and Dyatlov--Wang--Zworski, which gave a description of singular profiles when the underlying dynamics has periodic attractors. Our result is proved by showing the spectral measure near the forcing frequency is very small., Comment: 67 pages, 4 figures

We establish the local-to-global property of the synthetic curvature-dimension condition for essentially non-branching locally finite metric-measure spaces, extending the work [F. Cavalletti, E. Milman \textit{Invent. Math.} 226 (2021), no. 1, 1-137]., Comment: 21 pages. The definition of CD^1(K,N) is changed to a more convenient one. The part on the implication from CD^1 to CD is extensively expanded

The Poincar\'e problem is a model of two-dimensional internal waves in stable-stratified fluid. The chess billiard flow, a variation of a typical billiard flow, drives the formation behind and describes the evolution of these internal waves, and its trajectories can be represented as rotations around the boundary of a given domain. We find that for sufficiently irrational rotation in the square, or when the rotation number $r(\lambda)$ is Diophantine, the regularity of the solution $u(t)$ of the evolution problem correlates directly to the regularity of the forcing function $f(x)$. Additionally, we show that when $f$ is smooth, then $u$ is also smooth. These results extend studies that have examined singularity points, or the lack of regularity, in rational rotations of the chess billiard flow. We also present numerical simulations in various geometries that analyze plateau formation and fractal dimension in $r(\lambda)$ and conjecture an extension of our results. Our results can be applied in modeling two dimensional oceanic waves, and they also relate the classical quantum correspondence to fluid study.

We extend the result of Lisini (Calc Var Partial Differ Equ 28:85-120, 2007) on the superposition principle for absolutely continuous curves in $p$-Wasserstein spaces to the special case of $p=1$. In contrast to the case of $p>1$, it is not always possible to have lifts on absolutely continuous curves. Therefore, one needs to relax the notion of a lift by considering curves of bounded variation, or shortly BV-curves, and replace the metric speed by the total variation measure. We prove that any BV-curve in a 1-Wasserstein space can be represented by a probability measure on the space of BV-curves which encodes the total variation measure of the Wasserstein curve. In particular, when the curve is absolutely continuous, the result gives a lift concentrated on BV-curves which also characterizes the metric speed. The main theorem is then applied for the characterization of geodesics and the study of the continuity equation in a discrete setting., Comment: 37 pages, 3 figures

It is well established that graph neural networks (GNNs) can be interpreted and designed from the perspective of optimization objective. With this clear optimization objective, the deduced GNNs architecture has sound theoretical foundation, which is able to flexibly remedy the weakness of GNNs. However, this optimization objective is only proved for GNNs with single-relational graph. Can we infer a new type of GNNs for multi-relational graphs by extending this optimization objective, so as to simultaneously solve the issues in previous multi-relational GNNs, e.g., over-parameterization? In this paper, we propose a novel ensemble multi-relational GNNs by designing an ensemble multi-relational (EMR) optimization objective. This EMR optimization objective is able to derive an iterative updating rule, which can be formalized as an ensemble message passing (EnMP) layer with multi-relations. We further analyze the nice properties of EnMP layer, e.g., the relationship with multi-relational personalized PageRank. Finally, a new multi-relational GNNs which well alleviate the over-smoothing and over-parameterization issues are proposed. Extensive experiments conducted on four benchmark datasets well demonstrate the effectiveness of the proposed model., Comment: 7 pages, 3 figures, published to IJCAI 2022

We find Weyl upper bounds for the quantum open baker's map in the semiclassical limit. For the number of eigenvalues in an annulus, we derive the asymptotic upper bound $\mathcal O(N^\delta)$ where $\delta$ is the dimension of the trapped set of the baker's map and $(2 \pi N)^{-1}$ is the semiclassical parameter, which improves upon the previous result of $\mathcal O(N^{\delta + \epsilon})$. Furthermore, we derive a Weyl upper bound with explicit dependence on the inner radius of the annulus for quantum open baker's maps with Gevrey cutoffs., Comment: 24 pages, 5 figures

Data augmentation is an important component in the robustness evaluation of models in natural language processing (NLP) and in enhancing the diversity of the data they are trained on. In this paper, we present NL-Augmenter, a new participatory Python-based natural language augmentation framework which supports the creation of both transformations (modifications to the data) and filters (data splits according to specific features). We describe the framework and an initial set of 117 transformations and 23 filters for a variety of natural language tasks. We demonstrate the efficacy of NL-Augmenter by using several of its transformations to analyze the robustness of popular natural language models. The infrastructure, datacards and robustness analysis results are available publicly on the NL-Augmenter repository (https://github.com/GEM-benchmark/NL-Augmenter)., Comment: 39 pages, repository at https://github.com/GEM-benchmark/NL-Augmenter

The safety of connected automated vehicles (CAVs) relies on the reliable and efficient raw data sharing from multiple types of sensors. The 5G millimeter wave (mmWave) communication technology can enhance the environment sensing ability of different isolated vehicles. In this paper, a joint sensing and communication integrated system (JSCIS) is designed to support the dynamic frame structure configuration for sensing and communication dual functions based on the 5G New Radio protocol in the mmWave frequency band, which can solve the low latency and high data rate problems of raw sensing data sharing among CAVs. To evaluate the timeliness of raw sensing data transmission, the best time duration allocation ratio of sensing and communication dual functions for one vehicle is achieved by modeling the M/M/1 queuing problem using the age of information (AoI) in this paper. Furthermore, the resource allocation optimization problem among multiple CAVs is formulated as a non-cooperative game using the radar mutual information as a key indicator. And the feasibility and existence of pure strategy Nash equilibrium (NE) are proved theoretically, and a centralized time resource allocation (CTRA) algorithm is proposed to achieve the best feasible pure strategy NE. Finally, both simulation and hardware testbed are designed, and the results show that the proposed CTRA algorithm can improve the radar total mutual information by 26%, and the feasibility of the proposed JSCIS is achieved with an acceptable radar ranging accuracy within 0.25 m, as well as a stable data rate of 2.8 Gbps using the 28 GHz mmWave frequency band.

In this paper, we focus on studying duplicate logging statements, which are logging statements that have the same static text message. We manually studied over 4K duplicate logging statements and their surrounding code in five large-scale open source systems. We uncovered five patterns of duplicate logging code smells. For each instance of the duplicate logging code smell, we further manually identify the potentially problematic and justifiable cases. Then, we contact developers to verify our manual study result. We integrated our manual study result and the feedback of developers into our automated static analysis tool, DLFinder, which automatically detects problematic duplicate logging code smells. We evaluated DLFinder on the five manually studied systems and three additional systems. In total, combining the results of DLFinder and our manual analysis, we reported 91 problematic duplicate logging code smell instances to developers and all of them have been fixed. We further study the relationship between duplicate logging statements, including the problematic instances of duplicate logging code smells, and code clones. We find that 83% of the duplicate logging code smell instances reside in cloned code, but 17% of them reside in micro-clones that are difficult to detect using automated clone detection tools. We also find that more than half of the duplicate logging statements reside in cloned code snippets, and a large portion of them reside in very short code blocks which may not be effectively detected by existing code clone detection tools. Our study shows that, in addition to general source code that implements the business logic, code clones may also result in bad logging practices that could increase maintenance difficulties., Comment: Accepted at IEEE Transactions on Software Engineering

Neural Machine Translation models are sensitive to noise in the input texts, such as misspelled words and ungrammatical constructions. Existing robustness techniques generally fail when faced with unseen types of noise and their performance degrades on clean texts. In this paper, we focus on three types of realistic noise that are commonly generated by humans and introduce the idea of visual context to improve translation robustness for noisy texts. In addition, we describe a novel error correction training regime that can be used as an auxiliary task to further improve translation robustness. Experiments on English-French and English-German translation show that both multimodal and error correction components improve model robustness to noisy texts, while still retaining translation quality on clean texts., Comment: Accepted at Findings of EMNLP 2021; add acknowledgements

We prove the uniqueness of several excited states to the ODE $\ddot y(t) + \frac{2}{t} \dot y(t) + f(y(t)) = 0$, $y(0) = b$, and $\dot y(0) = 0$ for the model nonlinearity $f(y) = y^3 - y$. The $n$-th excited state is a solution with exactly $n$ zeros and which tends to $0$ as $t \to \infty$. These represent all smooth radial nonzero solutions to the PDE $\Delta u + f(u)= 0$ in $H^1$. We interpret the ODE as a damped oscillator governed by a double-well potential, and the result is proved via rigorous numerical analysis of the energy and variation of the solutions. More specifically, the problem of uniqueness can be formulated entirely in terms of inequalities on the solutions and their variation, and these inequalities can be verified numerically., Comment: 19 pages, 5 figures. Published in Analysis & PDE