SmartGift: Learning to Generate Practical Inputs for Testing Smart Contracts

Zhou, Teng; Liu, Kui; Li, Li; Liu, Zhe; Klein, Jacques; Bissyandé, Tegawendé F.

doi:10.1109/icsme52107.2021.00009

Cited by 14 publications

(7 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[32], [37]- [39], [42], [44], [45], [49] ULM Unidirectional Language Model A left-to-right language modeling task, asking the model to guess one masked token in an instance by only considering the leftward tokens (i.e., the tokens preceding the masked one).…”

Section: A Transformer Modelmentioning

confidence: 99%

Automating Code-Related Tasks Through Transformers: The Impact of Pre-training

Tufano¹,

Pascarella²,

Bavota³

2023

Preprint

View full text Add to dashboard Cite

Transformers have gained popularity in the software engineering (SE) literature. These deep learning models are usually pre-trained through a self-supervised objective, meant to provide the model with basic knowledge about a language of interest (e.g., Java). A classic pre-training objective is the masked language model (MLM), in which a percentage of tokens from the input (e.g., a Java method) is masked, with the model in charge of predicting them. Once pre-trained, the model is then finetuned to support the specific downstream task of interest (e.g., code summarization). While there is evidence suggesting the boost in performance provided by pre-training, little is known about the impact of the specific pre-training objective(s) used. Indeed, MLM is just one of the possible pre-training objectives and recent work from the natural language processing field suggest that pre-training objectives tailored for the specific downstream task of interest may substantially boost the model's performance. For example, in the case of code summarization, a tailored pretraining objective could be the identification of an appropriate name for a given method, considering the method name to generate as an extreme summary. In this study, we focus on the impact of pre-training objectives on the performance of transformers when automating code-related tasks. We start with a systematic literature review aimed at identifying the pre-training objectives used in SE. Then, we pre-train 32 transformers using both (i) generic pre-training objectives usually adopted in SE; and (ii) pre-training objectives tailored to specific code-related tasks subject of our experimentation, namely bug-fixing, code summarization, and code completion. We also compare the pretrained models with non pre-trained ones and show the advantage brought by pre-training in different scenarios, in which more or less fine-tuning data are available. Our results show that: (i) pre-training helps in boosting performance only if the amount of fine-tuning data available is small; (ii) the MLM objective is usually sufficient to maximize the prediction performance of the model, even when comparing it with pre-training objectives specialized for the downstream task at hand.

show abstract

Section: A Transformer Modelmentioning

confidence: 99%

Automating Code-Related Tasks Through Transformers: The Impact of Pre-training

Tufano¹,

Pascarella²,

Bavota³

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…If the gas allocated by the sender contract is insufficient to execute a costly fallback function, it sends an out-of-gas exception to the sender contract. If the sender does not check the exception properly, it will not realize the unsuccessful transfer, resulting in a gasless send [102,148,239].…”

Section: Division-by-zero Guardingmentioning

confidence: 99%

“…If a user-provided (nontrusted) value flows into the contract address invoked through a delegatecall, a nontrusted contract could be invoked, resulting in a high-risk security vulnerability called tainted or dangerous delegatecall [14,38,102,216,239]. This can be addressed by establishing guarding of delegatecall, which asserts that functions with delegatecall are either private or accessible from public functions through specific checks only and if the target address in a delegatecall is derived from user-provided input, it is checked against trusted contract addresses [52], and it explicitly flows into arguments of a delegatecall instruction [93].…”

Section: Information Flow-relatedmentioning

confidence: 99%

“…Fuzzing has been used to check user-specified custom properties and Assertions [81,204]. Other properties addressed using Fuzzing are two language constructs-related properties, namely, Guarding of selfdestruct (Section Ca) [204] and Exception Handling (Section 5.4.7) [93,102,136,148,204,210,216,239]. Fuzzing has also been used to check fundamental as well as complex domain-specific properties, as shown in table 5.…”

Section: Concolicmentioning

confidence: 99%

See 1 more Smart Citation

Pre-deployment Analysis of Smart Contracts -- A Survey

Munir¹,

Taha²

2023

Preprint

View full text Add to dashboard Cite

Smart contracts are programs that execute transactions involving independent parties and cryptocurrencies. As programs, smart contracts are susceptible to a wide range of errors and vulnerabilities. Such vulnerabilities can result in significant losses. Furthermore, by design, smart contract transactions are irreversible. This creates a need for methods to ensure the correctness and security of contracts pre-deployment. Recently there has been substantial research into such methods. The sheer volume of this research makes articulating state-of-the-art a substantial undertaking. To address this challenge, we present a systematic review of the literature. A key feature of our presentation is to factor out the relationship between vulnerabilities and methods through properties. Specifically, we enumerate and classify smart contract vulnerabilities and methods by the properties they address. The methods considered include static analysis as well as dynamic analysis methods and machine learning algorithms that analyze smart contracts before deployment.Several patterns about the strengths of different methods emerge through this classification process.CCS Concepts: • General and reference → Surveys and overviews; • Software and its engineering → Automated static analysis; Dynamic analysis; Software verification.

show abstract

“…In the papers [17], [19], [29], [30], [31], [34], [45], and [61], a newer method called mutation testing was introduced for the specific purpose of testing smart contracts. In addition, in the papers [20], [21], [23], [24], [26], [41], [42], [48], [53], [54], [55], and [56], another method called fuzz testing was introduced as well.…”

Section: ) Testing Data Challengesmentioning

confidence: 99%

Systematic Mapping of Testing Smart Contracts for Blockchain Applications

Imperius

Alahmar

2022

IEEE Access

View full text Add to dashboard Cite

In the last few years, the technological future becoming apparent by the introduction of smart contracts into mainstream technology, specifically in the development of Web3 and the metaverse. Smart contracts will play a vital role in the decentralization and autonomy of the day-to-day tasks that must be completed. Several literature reviews, considered secondary sources, highlight the current state of testing methods for smart contracts made for Blockchain applications. In this paper, we present the results from a systematic mapping study to give structure to the information found from primary sources. Systematic mapping is a well-known method to identify and categorize research papers in a field with an increasing amount of literature. For this systematic mapping, we searched for studies between 2017 and present-day (March 2022) and were able to find 303 results, from which 47 were selected, by specific inclusion and exclusion criteria, to be relevant to this study. A concept map was created from the information gathered from primary sources to the attributes such as research type, contribution type, blockchain network, smart contract language, development process, testing methods, and testing environment. We also categorized the trends and demographics found in the selected papers based on publication year, author's country, and more. The results of this systematic mapping showed that this field is very new and quickly increasing with new research. The researchers that are interested in this field could use the results found to create opportunities for their future work.

show abstract

SmartGift: Learning to Generate Practical Inputs for Testing Smart Contracts

Cited by 14 publications

References 37 publications

Automating Code-Related Tasks Through Transformers: The Impact of Pre-training

Automating Code-Related Tasks Through Transformers: The Impact of Pre-training

Pre-deployment Analysis of Smart Contracts -- A Survey

Systematic Mapping of Testing Smart Contracts for Blockchain Applications

Contact Info

Product

Resources

About