To reduce the risks of malicious software, malware detection methods using machine learning have received tremendous attention in recent years. Most conventional are based on supervised learning, which relies static features with definite labels. However, studies shown models vulnerable to deliberate attacks. This work tends expose and demonstrate weakness these models. A DQEAF framework reinforcement evade anti-malware engines is presented. trains an AI agent through a neural network by...

Bitcoin, one of the major cryptocurrencies, presents great opportunities and challenges with its tremendous potential returns accompanying high risks. The volatility Bitcoin complex factors affecting them make study effective price forecasting methods practical importance to financial investors researchers worldwide. In this paper, we propose a novel approach called MRC-LSTM, which combines Multi-scale Residual Convolutional neural network (MRC) Long Short-Term Memory (LSTM) implement...

Due to the completely open-source nature of Android, exploitable vulnerability malware attacks is increasing. To stay ahead other similar review work attempting deal with serious security problem Android environment, this not only summarizes approaches in classification phase but also lays emphasis on feature selection algorithm and presents some areas neglected previous works field detection, like limitations commonly applied datasets machine learning-based models. In paper, OS selection,...

With the growing popularity of information digitization and advancement executable file detection technology, PDF has emerged as an important carrier malicious documents. Despite improved efficacy machine learning-based classifiers in detecting malware, adversaries have proposed a variety countermeasures to evade detection, such generating adversarial examples. In contrast other peer works attempting expose vulnerability models, this work addresses deficiencies existing research by pointing...

Source code vulnerabilities are one of the significant threats to software security. Existing deep learning-based detection methods have proven their effectiveness. However, most them extract information on a single intermediate representation (IRC), which often fails multiple hidden in fully, significantly limiting performance. To address this problem, we propose VulMPFF, vulnerability method that fuses features under perspectives. It extracts IRC from three perspectives: sequence, lexical...

Vulnerability detection on source code can prevent the risk of cyber-attacks as early possible. However, lacking fine-grained analysis has rendered existing solutions still suffering from low performance; besides, explosive growth open-source projects dramatically increased complexity and diversity code. This paper presents HGVul, a vulnerability method based heterogeneous intermediate representation The key proposed is handling source-level (SIR) without expert knowledge. It first extracts...

With the large-scale development of open source software, software plagiarism has become a serious threat to industry and intellectual property. As latest technique detection, dynamic birthmark attracted much attention in recent years. Most existing birthmarks focus on how resist obfuscation techniques such as compiler optimizations strong obfuscations implemented tools. However, they pay little packers, especially encryption packer which is commonly used protection well plagiarism. When...

Due to the completely open-source nature of Android, exploitable vulnerability malware attacks is increasing. Machine learning has led a great evolution in Android detection recent years, which typically applied classification phase. Applying neural networks feature selection phase another topic worth investigating, while correlation between features could be ignored some traditional ranking-based algorithms. And it time-consuming for exploring all possible valid subsets when processing...

Third-party library (TPL) reuse may introduce vulnerable or malicious code and expose the software, which exposes them to potential risks. Thus, it is essential identify third-party dependencies take immediate corrective action fix critical vulnerabilities when a damaged reusable component found reported. However, most of existing methods only rely on syntactic features, results in low recognition accuracy significantly discounts detection performance by obfuscation techniques. In addition,...

Engineering equipment can efficiently assist humans in completing engineering projects and reduce work intensity. As the difficulty scope of operations increase, higher requirements are placed on function integration. In product design, excessive user design constraints lead to discrete results, difficult converge. Under multi-functionality high integration constraints, traditional mapping sub-functions structures one by results overabundant mappings, discrete, redundant after mapping,...

Abstract China is a country prone to geological hazards, and it has become especially urgent improve citizens’ awareness of earthquake safety prevention escape ability in the effort protect ourselves from disaster. However, effect current popular science education, there still significant lack skills masses emergency risk avoidance self-rescue. Based on Unity3D game engine, we developed an EES (Earthquake Escape Simulator) disaster knowledge system, training platform for primary secondary...

Due to the completely open-source nature of Android, exploitable vulnerability malware attacks is increasing. Machine learning, leading a great evolution in Android detection recent years, typically applied classification phase. Since correlation between features ignored some traditional ranking-based feature selection algorithms, applying wrapper-based models topic worth investigating. Though considering features, approaches are time-consuming for exploring all possible valid subsets when...

Recently, many machine learning methods had been proposed to detect malware, and they were proved have a better performance on polymorphic malware families. However, it has shown that are vulnerable mancraft adversarial examples. This paper proposes deep reinforcement framework for generating examples automatically. The method includes extracting bytes from benign executable files, appending them the end of or inserting into binary. model is evaluated with state-of-the-art classifier...