Large Language Models (LLMs) for code have gained significant attention recently. They can generate code in different programming languages based on provided prompts, fulfilling a long-lasting dream in Software Engineering (SE), i.e., automatic code generation. Similar to human-written code, LLM-generated code is prone to bugs, and these bugs have not yet been thoroughly examined by the community. Given the increasing adoption of LLM-based code generation tools (e.g., GitHub Copilot) in SE activities, it is critical to understand the characteristics of bugs contained in code generated by LLMs. This paper examines a sample of 333 bugs collected from code generated using three leading LLMs (i.e., CodeGen, PanGu-Coder, and Codex) and identifies the following 10 distinctive bug patterns: Misinterpretations, Syntax Error, Silly Mistake, Prompt-biased code, Missing Corner Case, Wrong Input Type, Hallucinated Object, Wrong Attribute, Incomplete Generation, and Non-Prompted Consideration. The bug patterns are presented in the form of a taxonomy. The identified bug patterns are validated using an online survey with 34 LLM practitioners and researchers. The surveyed participants generally asserted the significance and prevalence of the bug patterns. Researchers and practitioners can leverage these findings to develop effective quality assurance techniques for LLM-generated code. This study sheds light on the distinctive characteristics of LLM-generated code.
Large Language Models (LLMs) for code have gained significant attention recently. They can generate code in different programming languages based on provided prompts, fulfilling a long-lasting dream in Software Engineering (SE), i.e., automatic code generation. Similar to human-written code, LLM-generated code is prone to bugs, and these bugs have not yet been thoroughly examined by the community. Given the increasing adoption of LLM-based code generation tools (e.g., GitHub Copilot) in SE activities, it is critical to understand the characteristics of bugs contained in code generated by LLMs. This paper examines a sample of 333 bugs collected from code generated using three leading LLMs (i.e., CodeGen, PanGu-Coder, and Codex) and identifies the following 10 distinctive bug patterns: Misinterpretations, Syntax Error, Silly Mistake, Prompt-biased code, Missing Corner Case, Wrong Input Type, Hallucinated Object, Wrong Attribute, Incomplete Generation, and Non-Prompted Consideration. The bug patterns are presented in the form of a taxonomy. The identified bug patterns are validated using an online survey with 34 LLM practitioners and researchers. The surveyed participants generally asserted the significance and prevalence of the bug patterns. Researchers and practitioners can leverage these findings to develop effective quality assurance techniques for LLM-generated code. This study sheds light on the distinctive characteristics of LLM-generated code.
Code auditing ensures that the developed code adheres to standards, regulations, and copyright protection by verifying that it does not contain code from protected sources. The recent advent of Large Language Models (LLMs) as coding assistants in the software development process poses new challenges for code auditing. The dataset for training these models is mainly collected from publicly available sources. This raises the issue of intellectual property infringement as developers' codes are already included in the dataset. Therefore, auditing code developed using LLMs is challenging, as it is difficult to reliably assert if an LLM used during development has been trained on specific copyrighted codes, given that we do not have access to the training datasets of these models. Given the non-disclosure of the training datasets, traditional approaches such as code clone detection are insufficient for asserting copyright infringement. To address this challenge, we propose a new approach, TraWiC; a model-agnostic and interpretable method based on membership inference for detecting code inclusion in an LLM's training dataset. We extract syntactic and semantic identifiers unique to each program to train a classifier for detecting code inclusion. In our experiments, we observe that TraWiC is capable of detecting 83.87% of codes that were used to train an LLM. In comparison, the prevalent clone detection tool NiCad is only capable of detecting 47.64%. In addition to its remarkable performance, TraWiC has low resource overhead in contrast to pair-wise clone detection that is conducted during the auditing process of tools like CodeWhisperer reference tracker, across thousands of code snippets.
Since its release in November 2022, ChatGPT has shaken up Stack Overflow, the premier platform for developers' queries on programming and software development. Demonstrating an ability to generate instant, human-like responses to technical questions, ChatGPT has ignited debates within the developer community about the evolving role of human-driven platforms in the age of generative AI. Two months after ChatGPT's release, Meta released its answer with its own Large Language Model (LLM) called LLaMA: the race was on. We conducted an empirical study analyzing questions from Stack Overflow and using these LLMs to address them. This way, we aim to (ii) measure user engagement evolution with Stack Overflow over time; (ii) quantify the reliability of LLMs' answers and their potential to replace Stack Overflow in the long term; (iii) identify and understand why LLMs fails; and (iv) compare LLMs together. Our empirical results are unequivocal: ChatGPT and LLaMA challenge human expertise, yet do not outperform it for some domains, while a significant decline in user posting activity has been observed. Furthermore, we also discuss the impact of our findings regarding the usage and development of new LLMs.
The ubiquity of large-scale Pre-Trained Models (PTMs) is on the rise, sparking interest in model hubs, and dedicated platforms for hosting PTMs. Despite this trend, a comprehensive exploration of the challenges that users encounter and how the community leverages PTMs remains lacking. To address this gap, we conducted an extensive mixed-methods empirical study by focusing on discussion forums and the model hub of HuggingFace, the largest public model hub. Based on our qualitative analysis, we present a taxonomy of the challenges and benefits associated with PTM reuse within this community. We then conduct a quantitative study to track model-type trends and model documentation evolution over time. Our findings highlight prevalent challenges such as limited guidance for beginner users, struggles with model output comprehensibility in training or inference, and a lack of model understanding. We also identified interesting trends among models where some models maintain high upload rates despite a decline in topics related to them. Additionally, we found that despite the introduction of model documentation tools, its quantity has not increased over time, leading to difficulties in model comprehension and selection among users. Our study sheds light on new challenges in reusing PTMs that were not reported before and we provide recommendations for various stakeholders involved in PTM reuse.
Context: Deep learning has achieved remarkable progress in various domains. However, like traditional software systems, deep learning systems contain bugs, which can have severe impacts, as evidenced by crashes involving autonomous vehicles. Despite substantial advancements in deep learning techniques, little research has focused on reproducing deep learning bugs, which hinders resolving them. Existing literature suggests that only 3% of deep learning bugs are reproducible, underscoring the need for further research. Objective: This paper examines the reproducibility of deep learning bugs. We identify edit actions and useful information that could improve deep learning bug reproducibility. Method: First, we construct a dataset of 668 deep learning bugs from Stack Overflow and Defects4ML across 3 frameworks and 22 architectures. Second, we select 102 bugs using stratified sampling and try to determine their reproducibility. While reproducing these bugs, we identify edit actions and useful information necessary for their reproduction. Third, we used the Apriori algorithm to identify useful information and edit actions required to reproduce specific bug types. Finally, we conduct a user study with 22 developers to assess the effectiveness of our findings in real-life settings. Results: We successfully reproduced 85 bugs and identified ten edit actions and five useful information categories that can help us reproduce deep learning bugs. Our findings improved bug reproducibility by 22.92% and reduced reproduction time by 24.35% based on our user study. Conclusions: Our research addresses the critical issue of deep learning bug reproducibility. Practitioners and researchers can leverage our findings to improve deep learning bug reproducibility.
One goal of technical online communities is to help developers find the right answer in one place. A single question can be asked in different ways with different wordings, leading to the existence of duplicate posts on technical forums. The question of how to discover and link duplicate posts has garnered the attention of both developer communities and researchers. For example, Stack Overflow adopts a voting-based mechanism to mark and close duplicate posts. However, addressing these constantly emerging duplicate posts in a timely manner continues to pose challenges. Therefore, various approaches have been proposed to detect duplicate posts on technical forum posts automatically. The existing methods suffer from limitations either due to their reliance on handcrafted similarity metrics which can not sufficiently capture the semantics of posts, or their lack of supervision to improve the performance. Additionally, the efficiency of these methods is hindered by their dependence on pair-wise feature generation, which can be impractical for large amount of data. In this work, we attempt to employ and refine the GPT-3 embeddings for the duplicate detection task. We assume that the GPT-3 embeddings can accurately represent the semantics of the posts. In addition, by training a Siamese-based network based on the GPT-3 embeddings, we obtain a latent embedding that accurately captures the duplicate relation in technical forum posts. Our experiment on a benchmark dataset confirms the effectiveness of our approach and demonstrates superior performance compared to baseline methods. When applied to the dataset we constructed with a recent Stack Overflow dump, our approach attains a Top-1, Top-5, and Top-30 accuracy of 23.1%, 43.9%, and 68.9%, respectively. With a manual study, we confirm our approach's potential of finding unlabelled duplicates on technical forums.
With the rapid growth of the developer community, the amount of posts on online technical forums has been growing rapidly, which poses difficulties for users to filter useful posts and find important information. Tags provide a concise feature dimension for users to locate their interested posts and for search engines to index the most relevant posts according to the queries. However, most tags are only focused on the technical perspective (e.g., program language, platform, tool). In most cases, forum posts in online developer communities reveal the author's intentions to solve a problem, ask for advice, share information, etc. The modeling of the intentions of posts can provide an extra dimension to the current tag taxonomy. By referencing previous studies and learning from industrial perspectives, we create a refined taxonomy for the intentions of technical forum posts. Through manual labeling and analysis on a sampled post dataset extracted from online forums, we understand the relevance between the constitution of posts (code, error messages) and their intentions. Furthermore, inspired by our manual study, we design a pre-trained transformer-based model to automatically predict post intentions. The best variant of our intention prediction framework, which achieves a Micro F1-score of 0.589, Top 1-3 accuracy of 62.6% to 87.8%, and an average AUC of 0.787, outperforms the state-of-the-art baseline approach. Our characterization and automated classification of forum posts regarding their intentions may help forum maintainers or third-party tool developers improve the organization and retrieval of posts on technical forums. We have released our annotated dataset and codes in our supplementary material package.
Background: We are witnessing an increasing adoption of machine learning (ML), especially deep learning (DL) algorithms in many software systems, including safety-critical systems such as health care systems or autonomous driving vehicles. Ensuring the software quality of these systems is yet an open challenge for the research community, mainly due to the inductive nature of ML software systems. Traditionally, software systems were constructed deductively, by writing down the rules that govern the behavior of the system as program code. However, for ML software, these rules are inferred from training data. Few recent research advances in the quality assurance of ML systems have adapted different concepts from traditional software testing, such as mutation testing, to help improve the reliability of ML software systems. However, it is unclear if any of these proposed testing techniques from research are adopted in practice. There is little empirical evidence about the testing strategies of ML engineers. Aims: To fill this gap, we perform the first fine-grained empirical study on ML testing practices in the wild, to identify the ML properties being tested, the followed testing strategies, and their implementation throughout the ML workflow. Method: First, we systematically summarized the different testing strategies (e.g., Oracle Approximation), the tested ML properties (e.g., Correctness, Bias, and Fairness), and the testing methods (e.g., Unit test) from the literature. Then, we conducted a study to understand the practices of testing ML software. Results: In our findings: 1) we identified four (4) major categories of testing strategy including Grey-box, White-box, Black-box, and Heuristic-based techniques that are used by the ML engineers to find software bugs. 2) We identified 16 ML properties that are tested in the ML workflow.
As the demand for verifiability and testability of neural networks continues to rise, an increasing number of methods for generating test sets are being developed. However, each of these techniques tends to emphasize specific testing aspects and can be quite time-consuming. A straightforward solution to mitigate this issue is to transfer test sets between some benchmarked models and a new model under test, based on a desirable property one wishes to transfer. This paper introduces GIST (Generated Inputs Sets Transferability), a novel approach for the efficient transfer of test sets among Deep Learning models. Given a property of interest that a user wishes to transfer (e.g., coverage criterion), GIST enables the selection of good test sets from the point of view of this property among available ones from a benchmark. We empirically evaluate GIST on fault types coverage property with two modalities and different test set generation procedures to demonstrate the approach's feasibility. Experimental results show that GIST can select an effective test set for the given property to transfer it to the model under test. Our results suggest that GIST could be applied to transfer other properties and could generalize to different test sets' generation procedures and modalities