Clear

Wei, Moshi; Harzevili, Nima Shiri; Huang, Yuchao; Wang, Junjie; Wang, Song

doi:10.1145/3510003.3510159

Cited by 27 publications

(18 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, we fine-tune a pre-trained model CodeT5 [38] to learn the potential API completion patterns. Since most of the previous studies [13,16,32,40,50] mainly modeled API recommendation as the recommendation task, we first consider MRR (Mean Reciprocal Rank) [31] and MAP (Mean Average Precision) [34] as model performance evaluation measures. Moreover, since we model this problem as an automatic completion task in this study.…”

Section: Case1 Case2mentioning

confidence: 99%

“…Results. In our study, we first consider the classical API recommendation approaches (i,e., BIKER [13], RACK [32], and CLEAR [40]) as the baselines. Moreover, since we are the first to model this problem as the automatic completion task, we also consider pre-trained models (i,e., CodeBERT [8], UniXcoder [11] and PLBART [2]) as the baselines.…”

Section: Case1 Case2mentioning

confidence: 99%

“…Compare to previous studies on API recommendation [13,32,40], our study investigates a more practical problem. By providing some prompts with the query, our proposed approach APICom can help to provide accurate API completion suggestions, which can eventually generate APIs developers require.…”

Section: Case1 Case2mentioning

confidence: 99%

“…To compare the performance of APICom and the baselines, we first consider two performance measures: Mean Reciprocal Rank (MRR) [31] and Mean Average Precision (MAP) [34]. These two measures have also been widely used in previous API recommendation studies [13,32,40]. MRR.…”

Section: Evaluation Measuresmentioning

confidence: 99%

“…To demonstrate the competitiveness of our proposed approach APICom, we first compared APICom with state-of-the-art API recommendation baselines, such as BIKER [13], RACK [32], and CLEAR [40]. We introduce the characteristics of these baselines as follows.…”

Section: Baselinesmentioning

confidence: 99%

See 4 more Smart Citations

APICom: Automatic API Completion via Prompt Learning and Adversarial Training-based Data Augmentation

Gu,

Shen,

Chen

et al. 2023

Proceedings of the 14th Asia-Pacific Symposium on Internetware

View full text Add to dashboard Cite

Based on developer needs and usage scenarios, API (Application Programming Interface) recommendation is the process of assisting developers in finding the required API among numerous candidate APIs. Previous studies mainly modeled API recommendation as the recommendation task, which can recommend multiple candidate APIs for the given query, and developers may not yet be able to find what they need. Motivated by the neural machine translation research domain, we can model this problem as the generation task, which aims to directly generate the required API for the developer query. After our preliminary investigation, we find the performance of this intuitive approach is not promising. The reason is that there exists an error when generating the prefixes of the API. However, developers may know certain API prefix information during actual development in most cases. Therefore, we model this problem as the automatic completion task and propose a novel approach API-Com based on prompt learning, which can generate API related to the query according to the prompts (i.e., API prefix information). Moreover, the effectiveness of APICom highly depends on the quality of the training dataset. In this study, we further design a novel gradient-based adversarial training method ATCom for data augmentation, which can improve the normalized stability when generating adversarial examples. To evaluate the effectiveness of APICom, we consider a corpus of 33k developer queries and corresponding APIs. Compared with the state-of-the-art baselines, our experimental results show that APICom can outperform all baselines by at least 40.02%, 13.20%, and 16.31% in terms of the performance measures EM@1, MRR, and MAP. Finally, our ablation studies confirm the effectiveness of our component setting (such as our designed adversarial training method, our used pre-trained model, and prompt learning) in APICom.

show abstract

Section: Case1 Case2mentioning

confidence: 99%

Section: Case1 Case2mentioning

confidence: 99%

Section: Case1 Case2mentioning

confidence: 99%

Section: Evaluation Measuresmentioning

confidence: 99%

Section: Baselinesmentioning

confidence: 99%

See 3 more Smart Citations

APICom: Automatic API Completion via Prompt Learning and Adversarial Training-based Data Augmentation

Gu,

Shen,

Chen

et al. 2023

Proceedings of the 14th Asia-Pacific Symposium on Internetware

View full text Add to dashboard Cite

show abstract

Prompt enhance API recommendation: visualize the user’s real intention behind this query

Wang,

Chen,

Gao

et al. 2024

Autom Softw Eng

View full text Add to dashboard Cite

Developers frequently rely on APIs in their daily programming tasks, as APIs have become an indispensable tool for program development. However, with a vast number of open-source libraries available, selecting the appropriate API quickly can be a common challenge for programmers. Previous research on API recommendation primarily focused on designing better approaches to interpret user input. However, in practical applications, it is often difficult for users, especially novice programmers, to express their real intentions due to the limitations of language expression and programming capabilities. To address this issue, this paper introduces PTAPI, an approach that visualizes the user's real intentions based on their query to enhance recommendation performance. Firstly, PTAPI identifies the prompt template from Stack Overflow (SO) posts based on the user's input. Secondly, the obtained prompt template is combined with the user's input to generate a new question. Finally, the newly 1 generated question leverages dual information sources from SO posts and API official documentation to provide recommendations. To evaluate the effectiveness of PTAPI, we conducted experiments at both the class-level and method-level. The experimental results demonstrate the effectiveness of the proposed approach, with a significant improvement in the success rate.

show abstract