Junkai Chen, Huihui Huang, Yunbo Lyu, Junwen An, Jieke Shi, Chengran Yang, Ting Zhang, Haoye Tian, Yikun Li, Zhenhao Li, Xin Zhou, Xing Hu, David Lo. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Chengran Yang, Ting Zhang, Jinfeng Jiang, Xin Zhou, Haoye Tian, Mingzhe Du, Jieke Shi, Junkai Chen, Yikun Li, Eng Lieh Ouh, Lwin Khin Shar, David Lo. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

API documentation is often the most trusted resource for programming. Many approaches have been proposed to augment API documentation by summarizing complementary information from external resources like Stack Overflow. Existing extractive summarization approaches excel in producing faithful summaries that accurately represent the source content without input length restrictions. Nevertheless, they suffer from inherent readability limitations. On the other hand, our empirical study on the abstractive-based summarization method, i.e., GPT-4, reveals that GPT-4 can generate coherent and concise summaries but presents limitations in terms of informativeness and faithfulness. We introduce APIDOCBOOSTER, an extract-then-abstract framework that seamlessly fuses the advantages of both extractive (i.e., enabling faithful summaries without length limitation) and abstractive summarization (i.e., producing coherent and concise summaries). APIDocBooster consists of two stages: (1) Context-aware Sentence Section Classification (CSSC) and (2) UPdate SUMmarization (UPSUM). CSSC classifies APIrelevant information collected from multiple sources into API documentation sections. UPSUM generates extractive summaries distinct from original API documentation and then abstractive summaries guided by extractive summaries through in-context learning. To enable automatic evaluation, we construct the first dataset for API documentation augmentation. Our automatic evaluation results reveal that each stage in APIDocBooster outperforms its baselines by a large margin. Our human evaluation also demonstrates the superiority of APIDOCBOOSTER over GPT-4 and shows that it improves the informativeness, relevance and faithfulness by$\mathbf{1 6. 2 2 \%}, \mathbf{1 9. 4 4 \%}$, and$\mathbf{3 7. 1 4 \%}$, respectively.

Large language models (LLMs) have shown exceptional performance in code generation and understanding tasks, yet their high computational costs hinder broader adoption. One important factor is the inherent verbosity of programming languages, such as unnecessary formatting elements and lengthy boilerplate code. This leads to inflated token counts in both input and generated outputs, which increases inference costs and slows down the generation process. Prior work improves this through simplifying programming language grammar, reducing token usage across both code understanding and generation tasks. However, it is confined to syntactic transformations, leaving significant opportunities for token reduction unrealized at the semantic level.In this work, we propose Token Sugar, a concept that replaces frequent and verbose code patterns with reversible, token-efficient shorthand in the source code. To realize this concept in practice, we designed a systematic solution that mines high-frequency, token-heavy patterns from a code corpus, maps each to a unique shorthand, and integrates them into LLM pretraining via code transformation. With this solution, we obtain 799 (code pattern, shorthand) pairs, which can reduce up to 15.1% token count in the source code and is complementary to existing syntax-focused methods. We further trained three widely used LLMs on Token Sugar-augmented data. Experimental results show that these models not only achieve significant token savings (up to 11.2% reduction) during generation but also maintain near-identical Pass@1 scores compared to baselines trained on unprocessed code.

Sequential decision-making processes (SDPs) are fundamental for complex real-world challenges, such as autonomous driving, robotic control, and traffic management. While recent advances in Deep Learning (DL) have led to mature solutions for solving these complex problems, SDMs remain vulnerable to learning unsafe behaviors, posing significant risks in safety-critical applications. However, developing a testing framework for SDMs that can identify a diverse set of crash-triggering scenarios remains an open challenge. To address this, we propose CureFuzz, a novel curiosity-driven black-box fuzz testing approach for SDMs. CureFuzz proposes a curiosity mechanism that allows a fuzzer to effectively explore novel and diverse scenarios, leading to improved detection of crash-triggering scenarios. Additionally, we introduce a multi-objective seed selection technique to balance the exploration of novel scenarios and the generation of crash-triggering scenarios, thereby optimizing the fuzzing process. We evaluate CureFuzz on various SDMs and experimental results demonstrate that CureFuzz outperforms the state-of-the-art method by a substantial margin in the total number of faults and distinct types of crash-triggering scenarios. We also demonstrate that the crash-triggering scenarios found by CureFuzz can repair SDMs, highlighting CureFuzz as a valuable tool for testing SDMs and optimizing their performance.

With the increasing reliance on Open Source Software, users are exposed to third-party library vulnerabilities. Software Composition Analysis (SCA) tools have been created to alert users of such vulnerabilities. SCA requires the identification of vulnerability-fixing commits. Prior works have proposed methods that can automatically identify such vulnerability-fixing commits. However, identifying such commits is highly challenging, as only a very small minority of commits are vulnerability fixing. Moreover, code changes can be noisy and difficult to analyze. We observe that noise can occur at different levels of detail, making it challenging to detect vulnerability fixes accurately. To address these challenges and boost the effectiveness of prior works, we propose MiDas (Multi-Granularity Detector for Vulnerability Fixes). Unique from prior works, MiDas constructs different neural networks for each level of code change granularity, corresponding to commit-level, file-level, hunk-level, and line-level, following their natural organization. It then utilizes an ensemble model that combines all base models to generate the final prediction. This design allows MiDas to better handle the noisy and highly imbalanced nature of vulnerability-fixing commit data. Additionally, to reduce the human effort required to inspect code changes, we have designed an effort-aware adjustment for MiDas's outputs based on commit length. The evaluation results demonstrate that MiDas outperforms the current state-of-the-art baseline in terms of AUC by 4.9% and 13.7% on Java and Python-based datasets, respectively. Furthermore, in terms of two effort-aware metrics, EffortCost@L and Popt@L, MiDas also outperforms the state-of-the-art baseline, achieving improvements of up to 28.2% and 15.9% on Java, and 60% and 51.4% on Python, respectively.

Stack Overflow is often viewed as one of the most influential Software Question & Answer (SQA) websites, containing millions of programming-related questions and answers. Tags play a critical role in efficiently structuring the contents in Stack Overflow and are vital to support a range of site operations, e.g., querying relevant contents. Poorly selected tags often introduce extra noise and redundancy, which raises problems like tag synonym and tag explosion. Thus, an automated tag recommendation technique that can accurately recommend high-quality tags is desired to alleviate the problems mentioned above.

Prior studies have demonstrated that approaches to generate an answer summary for a given technical query in Software Question and Answer (SQA) sites are desired. We find that existing approaches are assessed solely through user studies. Hence, a new user study needs to be performed every time a new approach is introduced; this is time-consuming, slows down the development of the new approach, and results from different user studies may not be comparable to each other. There is a need for a benchmark with ground truth summaries as a complement assessment through user studies. Unfortunately, such a benchmark is non-existent for answer summarization for technical queries from SQA sites.

Programming videos on the Internet are valuable resources for learning programming skills. To find relevant videos, developers typically search online video platforms (e.g., YouTube) with keywords on topics they wish to learn. Developers often look for live-coding screencasts, in which the videos' authors perform live coding. Yet, not all programming videos are live-coding screencasts. In this work, we develop a tool named PSFinder to identify live-coding screencasts. PSFinder leverages a classifier to identify whether a video frame contains an IDE window. It uses a sampling strategy to pick a number of frames from an input video, runs the classifer on these frames, and then determines whether the video is a live-coding screencast based on frames classified as containing IDE window. In our preliminary experiment, PSFinder can effectively identify live-coding screencasts as it achieves an F1-score of 0.97.

APIs (Application Programming Interfaces) are reusable software libraries and are building blocks for modern rapid software development. Previous research shows that programmers frequently share and search for reviews of APIs on the mainstream software question and answer (Q&A) platforms like Stack Overflow, which motivates researchers to design tasks and approaches related to process API reviews automatically. Among these tasks, classifying API reviews into different aspects (e.g., performance or security), which is called the aspect-based API review classification, is of great importance. The current state-of-the-art (SOTA) solution to this task is based on the traditional machine learning algorithm. Inspired by the great success achieved by pre-trained models on many software engineering tasks, this study fine-tunes six pre-trained models for the aspect-based API review classification task and compares them with the current SOTA solution on an API review benchmark collected by Uddin et al. The investigated models include four models (BERT, RoBERTa, ALBERT and XLNet) that are pre-trained on natural languages, BERTOverflow that is pre-trained on text corpus extracted from posts on Stack Overflow, and CosSensBERT that is designed for handling imbalanced data. The results show that all the six fine-tuned models outperform the traditional machine learning-based tool. More specifically, the improvement on the F1-score ranges from 21.0% to 30.2%. We also find that BERTOverflow, a model pre-trained on the corpus from Stack Overflow, does not show better performance than BERT. The result also suggests that CosSensBERT also does not exhibit better performance than BERT in terms of F1, but it is still worthy of being considered as it achieves better performance on MCC and AUC.

Past studies have proposed solutions that analyze Stack Overflow content to help users find desired information or aid various downstream software engineering tasks. A common step performed by those solutions is to extract suitable representations of posts; typically, in the form of meaningful vectors. These vectors are then used for different tasks, for example, tag recommendation, relatedness prediction, post classification, and API recommendation. Intuitively, the quality of the vector representations of posts determines the effectiveness of the solutions in performing the respective tasks. In this work, to aid existing studies that analyze Stack Overflow posts, we propose a specialized deep learning architecture Post2Vec which extracts distributed representations of Stack Overflow posts. Post2Vec is aware of different types of content present in Stack Overflow posts, i.e., title, description, and code snippets, and integrates them seamlessly to learn post representations. Tags provided by Stack Overflow users that serve as a common vocabulary that captures the semantics of posts are used to guide Post2Vec in its task. To evaluate the quality of Post2Vec's deep learning architecture, we first investigate its end-to-end effectiveness in tag recommendation task. The results are compared to those of state-of-the-art tag recommendation approaches that also employ deep neural networks. We observe that Post2Vec achieves 15-25 percent improvement in terms of F1-score@5 at a lower computational cost. Moreover, to evaluate the value of representations learned by Post2Vec, we use them for three other tasks, i.e., relatedness prediction, post classification, and API recommendation. We demonstrate that the representations can be used to boost the effectiveness of state-of-the-art solutions for the three tasks by substantial margins (by 10, 7, and 10 percent in terms of F1-score, F1-score, and correctness, respectively). We release our replication package athttps://github.com/maxxbw/Post2Vec.