Understanding an unfamiliar codebase is an essential task for developers in various scenarios, such as during the onboarding process. Especially when the codebase is large and time is limited, achieving a decent level of comprehension remains challenging for both experienced and novice developers, even with the assistance of large language models (LLMs). Existing studies have shown that LLMs often fail to support users in understanding code structures or to provide user-centered, adaptive, and dynamic assistance in real-world settings.

Code review is an essential activity for ensuring the quality and maintainability of software projects. However, it is a time-consuming and often error-prone task that can significantly impact the development process. Recently, ChatGPT, a cutting-edge language model, has demonstrated impressive performance in various natural language processing tasks, suggesting its potential to automate code review processes. However, it is still unclear how well ChatGPT performs in code review tasks. To fill this gap, in this paper, we conduct the first empirical study to understand the capabilities of ChatGPT in code review tasks, specifically focusing on automated code refinement based on given code reviews. To conduct the study, we select the existing benchmark CodeReview and construct a new code review dataset with high quality. We use CodeReviewer, a state-of-the-art code review tool, as a baseline for comparison with ChatGPT. Our results show that ChatGPT outperforms CodeReviewer in code refinement tasks. Specifically, our results show that ChatGPT achieves higher EM and BLEU scores of 22.78 and 76.44 respectively, while the state-of-the-art method achieves only 15.50 and 62.88 on a high-quality code review dataset. We further identify the root causes for ChatGPT's underperformance and propose several strategies to mitigate these challenges. Our study provides insights into the potential of ChatGPT in automating the code review process, and highlights the potential research directions.

While AI-assisted individual qualitative analysis has been substantially studied, AI-assisted collaborative qualitative analysis (CQA) – a process that involves multiple researchers working together to interpret data—remains relatively unexplored. After identifying CQA practices and design opportunities through formative interviews, we designed and implemented CoAIcoder, a tool leveraging AI to enhance human-to-human collaboration within CQA through four distinct collaboration methods. With a between-subject design, we evaluated CoAIcoder with 32 pairs of CQA-trained participants across common CQA phases under each collaboration method. Our findings suggest that while using a shared AI model as a mediator among coders could improve CQA efficiency and foster agreement more quickly in the early coding stage, it might affect the final code diversity. We also emphasize the need to consider the independence level when using AI to assist human-to-human collaboration in various CQA scenarios. Lastly, we suggest design implications for future AI-assisted CQA systems.

Large garages are ubiquitous yet intricate scenes that present unique challenges due to their monotonous colors, repetitive patterns, reflective surfaces, and transparent vehicle glass. Conventional Structure from Motion (SfM) methods for camera pose estimation and 3D reconstruction often fail in these environments due to poor correspondence construction. To address these challenges, we introduce LetsGo, a LiDAR-assisted Gaussian splatting framework for large-scale garage modeling and rendering. We develop a handheld scanner, Polar, equipped with IMU, LiDAR, and a fisheye camera, to facilitate accurate data acquisition. Using this Polar device, we present the GarageWorld dataset, consisting of eight expansive garage scenes with diverse geometric structures, which will be made publicly available for further research. Our approach demonstrates that LiDAR point clouds collected by the Polar device significantly enhance a suite of 3D Gaussian splatting algorithms for garage scene modeling and rendering. We introduce a novel depth regularizer that effectively eliminates floating artifacts in rendered images. Additionally, we propose a multi-resolution 3D Gaussian representation designed for Level-of-Detail (LOD) rendering. This includes adapted scaling factors for individual levels and a random-resolution-level training scheme to optimize the Gaussians across different resolutions. This representation enables efficient rendering of large-scale garage scenes on lightweight devices via a web-based renderer. Experimental results on our GarageWorld dataset, as well as on ScanNet++ and KITTI-360, demonstrate the superiority of our method in terms of rendering quality and resource efficiency.

We present a novel Neural Refractive Field (NeReF) to recover wavefront of transparent fluids by simultaneously estimating the surface position and normal of the fluid front. Unlike prior arts that treat the reconstruction target as a single layer of the surface, NeReF is specifically formulated to recover a volumetric normal field with its corresponding density field. A query ray will be refracted by NeReF according to its accumulated refractive point and normal, and we employ the correspondences and uniqueness of refracted ray for NeReF optimization. We show NeReF, as a global optimization scheme, can more robustly tackle refraction distortions detrimental to traditional methods for correspondence matching. Furthermore, the continuous NeReF representation of wavefront enables view synthesis as well as normal integration. We validate our approach on both synthetic and real data and show it is particularly suitable for sparse multi-view acquisition. We hence build a small light field array and experiment on various surface shapes to demonstrate high fidelity NeReF reconstruction.

Machine learning (ML) enabled systems are emerging with recent breakthroughs in ML. A model-centric view is widely taken by the literature to focus only on the analysis of ML models. However, only a small body of work takes a system view that looks at how ML components work with the system and how they affect software engineering for ML-enabled systems. In this paper, we adopt this system view, and conduct a case study on Rasa 3.0, an industrial dialogue system that has been widely adopted by various companies around the world. Our goal is to characterize the complexity of such a large-scale ML-enabled system and to understand the impact of the complexity on testing. Our study reveals practical implications for software engineering for ML-enabled systems.

Fault diagnosis in robotic systems is challenging due to their complex and heterogeneous structures and complex interactions with physical environments. Given the complexities and uncertainties, we think it may be helpful to diagnose faults of a robotic system by understanding its behaviors from the perspective of observability. In this paper, we conduct an empirical study to explore the efficacy of combining different kinds of common observability data (i.e., logs, traces, and trajectories) for fault diagnosis in robotic systems. In the study, we investigate root causes of 398 bug cases in robotic systems to understand their characteristics. Furthermore, we replicate 23 bugs out of them and perform a fault diagnosis study in which participants diagnose each of the replicated bug with only observability data and record how useful observability data is. The bug case analysis study revealed that the root causes of bugs in robotic systems originate from various levels, including physical environment interaction (11.81%), hardware usage (14.82%), software implementation (49.25%), and system configuration (24.12%). The fault diagnosis study shows the combination of trace and trajectory data improves the fault diagnosis success rate by 58.33% and 8.33%, respectively, compared to using only logs. Our study promotes the vision of observability-based fault diagnosis in robotic systems.

Bugs and their fixes in the code evolution histories are important assets for many software engineering tasks such as deriving new state-of-the-art automatic bug fixing techniques. Existing bug datasets are either manually built which is difficult to grow efficiently to a scale large enough for massive data analysis, or lack of precise information of how bugs are introduced and fixed which is critical for in-depth analysis such as buggy/fixing code identification. Moreover, the types of the bugs are typically missing in the existing bug datasets, limiting the possibility of developing high-precision type-specific approaches for enterprise-level purposes. In this work, we propose BugMiner, an approach to automatically collecting bugs from code repositories by isolating the critical changes of the bugs. We also propose a learning-based approach for automating bug type classification with relatively small manual labels of bug types. We evaluate our approach regarding the precision of bug information and the efficiency of the bug-mining process with 2,082 bugs automatically mined from 100 open-source projects. We demonstrate the improved effectiveness and efficiency in bug-fixing location identification, compared to the SOTA BugBuilder, and high recall and precision in bug-inducing location identification. We also compare our learning-based bug classification approach to traditional baseline method, indicating about 17 % improvement in classification effectiveness under macro-F1.

Deep learning (DL) applications, built upon a heterogeneous and complex DL stack (e.g., Nvidia GPU, Linux, CUDA driver, Python runtime, and TensorFlow), are subject to software and hardware dependencies across the DL stack. One challenge in dependency management across the entire engineering lifecycle is posed by the asynchronous and radical evolution and the complex version constraints among dependencies. Developers may introduce dependency bugs (DBs) in selecting, using and maintaining dependencies. However, the characteristics of DBs in DL stack is still under-investigated, hindering practical solutions to dependency management in DL stack. To bridge this gap, this paper presents the first comprehensive study to characterize symptoms, root causes and fix patterns of DBs across the whole DL stack with 446 DBs collected from StackOverflow posts and GitHub issues. For each DB, we first investigate the symptom as well as the lifecycle stage and dependency where the symptom is exposed. Then, we analyze the root cause as well as the lifecycle stage and dependency where the root cause is introduced. Finally, we explore the fix pattern and the knowledge sources that are used to fix it. Our findings from this study shed light on practical implications on dependency management.

Deep learning (DL) has been widely applied to many domains. Unique challenges in engineering DL systems are posed by the programming paradigm shift from traditional systems to DL systems, and performance is one of the challenges. Performance problems (PPs) in DL systems can cause severe consequences such as excessive resource consumption and financial loss. While bugs in DL systems have been extensively investigated, PPs in DL systems have hardly been explored. To bridge this gap, we present the first comprehensive study to i) characterize symptoms, root causes, and introducing and exposing stages of PPs in DL systems developed in TensorFLow and Keras, with 224 PPs collected from 210 StackOverflow posts, and to ii) assess the capability of existing performance analysis approaches in tackling PPs, with a constructed benchmark of 58 PPs in DL systems. Our findings shed light on the implications on developing high-performance DL systems, and detecting and localizing PPs in DL systems. To demonstrate the usefulness of our findings, we develop a static checker DeepPerf to detect three types of PPs. It has detected 488 new PPs in 130 GitHub projects. 105 and 27 PPs have been confirmed and fixed.

IR-based fault localization approaches achieves promising results when locating faulty files by comparing a bug report with source code. Unfortunately, they become less effective to locate faulty methods. We conduct a preliminary study to explore its challenges, and identify three problems: the semantic gap problem, the representation sparseness problem, and the single revision problem.To tackle these problems, we propose MRAM, a mixed RNN and attention model, which combines bug-fixing features and method structured features to explore both implicit and explicit relevance between methods and bug reports for method level fault localization task. The core ideas of our model are: (1) constructing code revision graphs from code, commits and past bug reports, which reveal the latent relations among methods to augment short methods and as well provide all revisions of code and past fixes to train more accurate models; (2) embedding three method structured features (token sequences, API invocation sequences, and comments) jointly with RNN and soft attention to represent source methods and obtain their implicit relevance with bug reports; and (3) integrating multi-revision bug-fixing features, which provide the explicit relevance between bug reports and methods, to improve the performance.We have implemented MRAM and conducted a controlled experiment on five open-source projects. Comparing with state-of-the-art approaches, our MRAM improves MRR values by 3.8-5.1% (3.7-5.4%) when the dataset contains (does not contain) localized bug reports. Our statistics test shows that our improvements are significant.

Given a bug report of a project, the task of locating the faults of the bug report is called fault localization. To help programmers in the fault localization process, many approaches have been proposed, and have achieved promising results to locate faulty files. However, it is still challenging to locate faulty methods, because many methods are short and do not have sufficient details to determine whether they are faulty. In this paper, we present BugPecker, a novel approach to locate faulty methods based on its deep learning on revision graphs. Its key idea includes (1) building revision graphs and capturing the details of past fixes as much as possible, and (2) discovering relations inside our revision graphs to expand the details for methods and calculating various features to assist our ranking. We have implemented BugPecker, and evaluated it on three open source projects. The early results show that BugPecker achieves a mean average precision (MAP) of 0.263 and mean reciprocal rank (MRR) of 0.291, which improve the prior approaches significantly. For example, BugPecker improves the MAP values of all three projects by five times, compared with two recent approaches such as DNNLoc-m and BLIA 1.5.

Coding convention plays an important role in guaranteeing software quality. However, coding conventions are usually informally presented and inconvenient for programmers to use. In this paper, we present CocoQa, a system that answers programmer's questions about coding conventions. CocoQa answers questions by querying a knowledge graph for coding conventions. It employs 1) a subgraph matching algorithm that parses the question into a SPARQL query, and 2) a machine comprehension algorithm that uses an end-to-end neural network to detect answers from searched paragraphs. We have implemented CocoQa, and evaluated it on a coding convention QA dataset. The results show that CocoQa can answer questions about coding conventions precisely. In particular, CocoQa can achieve a precision of 82.92% and a recall of 91.10%. Repository: https://github.com/14dtj/CocoQa/ Video: https://youtu.be/VQaXi1WydAU.

Coding conventions are a set of coding guidelines used by software developers to improve the readability of source code, increase software maintainability, and promote the reuse of coding patterns.In this paper, we introduce CCBase, a knowledge base of coding conventions, that was constructed from online resources.Specifically, CCBase was constructed as follows.We designed the ontology of the coding convention domain, crawled data related to coding conventions from a variety of online resources, and then extracted entities and relations using an NLP-enabled rule matching method.To uncover the latent relations, we further proposed a similarity metric to reveal the similar-to and relate-to relations, and developed a RCE algorithm to establish a unified type hierarchy of coding conventions.The resulting knowledge base contains 3139 coding conventions for Java and C++, with 3761 entities and 767 relations.Furthermore, we have extended the usability of CCBase by developing a question answering system on the base.We have conducted experiments to evaluate CCBase.The experimental results show that CCBase has a wide coverage on entities and relations in coding conventions domain, and the QA system achieves an F1 score of 84.5% on 214 questions raised in StackOverflow.

With the trends of developing software on the Internet, many software crowdsourcing platforms are emerging.They attract a lot of developers to bid for crowdsourced projects and develop software systems collaboratively.In this paper, we present CrowDevBot, a task-oriented conversational bot for software crowdsourcing platform, that aims to assist online users in completing crowdsourcing-related tasks in a more natural manner.The key idea of CrowDevBot is to: (1) combine a rulebased method and an SVM-NaiveBayes-C4.5 integrated learning method to discover users' intention; (2) employ an integrated CRF (conditional random field) method with novel features to improve the performance of slot filling; and (3) leverage a software service knowledge base to unify entity names and predefine the key slots of user query.We implement CrowDevBot and integrate it into JointForce, an IT software crowdsourcing platform in China.To the best of our knowledge, this is the first time that a task-oriented conversational bot is practically used in software crowdsourcing platform(s).We evaluated our approach on real data set from JointForce.The results show that our intention detecting method achieves F1-score of 87% on the limited training data.For the slot filling, the F1-score of our integrated CRF model reaches 82%, 8% higher than that of the normal CRF model.