Search

Machine learning and its promising branch deep learning have proven to be effective in a wide range of application domains. Recently, several efforts have shown success in applying deep learning techniques for automatic vulnerability discovery, as alternatives to traditional static bug detection. In principle, these learning-based approaches are built on top of classification models using supervised learning. Depending on the different granularities to detect vulnerabilities, these approaches rely on learning models which are typically trained with well-labeled source code to predict whether a program method, a program slice, or a particular code line contains a vulnerability or not. The effectiveness of these models is normally evaluated against conventional metrics including precision, recall and F1 score. In this paper, we show that despite yielding promising numbers, the above evaluation strategy can be insufficient and even misleading when evaluating the effectiveness of current learning-based approaches. This is because the underlying learning models only produce the classification results or report individual/isolated program statements, but are unable to pinpoint bug-triggering paths, which is an effective way for bug fixing and the main aim of static bug detection. Our key insight is that a program method or statement can only be stated as vulnerable in the context of a bug-triggering path. In this work, we systematically study the gap between recent learning-based approaches and conventional static bug detectors in terms of fine-grained metrics called BTP metrics using bug-triggering paths. We then characterize and compare the quality of the prediction results of existing learning-based detectors under different granularities. Finally, our comprehensive empirical study reveals several key issues and challenges in developing classification models to pinpoint bug-triggering paths and calls for more advanced learning-based bug detection techniques.

Cross-language programming is a common practice within the software development industry, offering developers a multitude of advantages such as expressiveness, interoperability, and cross-platform compatibility, for developing large-scale applications. As an important example, JNI (Java Native Interface) programming is widely used in diverse scenarios where Java interacts with code written in other programming languages, such as C or C++. Conventional static analysis based on a single programming language faces challenges when it comes to tracing the flow of values across multiple modules that are coded in different programming languages. In this paper, we introduce CSS, a new Caller-Sensitive Specification approach designed to enhance the static taint analysis of Java programs employing JNI to interface with C/C++ code. In contrast to conservative specifications, this approach takes into consideration the calling context of the invoked C/C++ functions (or cross-language context), resulting in more precise and concise specifications for the side effects of native code. Furthermore, CSS specifically enhances the capabilities of Java analyzers, enabling them to perform precise static taint analysis across language boundaries into native code. The experimental results show that CSS can accurately summarize value-flow information and enhance the ability of Java monolingual static analyzers for cross-language taint flow tracking.

Deep neural networks (DNNs) have widespread applications in industries such as image recognition, supply chain, medical diagnosis, and autonomous driving. However, previous work has shown that the high accuracy of a DNN model does not imply high robustness (i.e., consistent performances on new and future datasets) because the input data and external environment (e.g., software and model configurations) for a deployed model are constantly changing. Therefore, ensuring robustness is crucial to enhance business and consumer confidence. Previous research focuses mostly on the data aspect of model variance. This article takes a holistic view of DNN robustness by summarizing the issues related to both data and software configuration variances. We also present a predictive framework using search-based optimization to generate representative variances for robust learning, considering data and configurations.

Many machine learning applications encounter situations where model providers are required to further refine the previously trained model so as to gratify the specific need of local users. This problem is reduced to the standard model tuning paradigm if the target data is permissibly fed to the model. However, it is rather difficult in a wide range of practical cases where target data is not shared with model providers but commonly some evaluations about the model are accessible. In this paper, we formally set up a challenge named Earning eXtra PerformancE from restriCTive feEDdbacks (EXPECTED) to describe this form of model tuning problems. Concretely, EXPECTED admits a model provider to access the operational performance of the candidate model multiple times via feedback from a local user (or a group of users). The goal of the model provider is to eventually deliver a satisfactory model to the local user(s) by utilizing the feedbacks. Unlike existing model tuning methods where the target data is always ready for calculating model gradients, the model providers in EXPECTED only see some feedbacks which could be as simple as scalars, such as inference accuracy or usage rate. To enable tuning in this restrictive circumstance, we propose to characterize the geometry of the model performance with regard to model parameters through exploring the parameters’ distribution. In particular, for deep models whose parameters distribute across multiple layers, a more query-efficient algorithm is further tailor-designed that conducts layerwise tuning with more attention to those layers which pay off better. Our theoretical analyses justify the proposed algorithms from the aspects of both efficacy and efficiency. Extensive experiments on different applications demonstrate that our work forges a sound solution to the EXPECTED problem, which establishes the foundation for future studies towards this direction.

Software clone detection identifies similar or identical code snippets. It has been an active research topic that attracts extensive attention over the last two decades. In recent years, machine learning (ML) based detectors, especially deep learning-based ones, have demonstrated impressive capability on clone detection. It seems that this longstanding problem has already been tamed owing to the advances in ML techniques. In this work, we would like to challenge the robustness of the recent ML-based clone detectors through code semantic-preserving transformations. We first utilize fifteen simple code transformation operators combined with commonly-used heuristics (i.e., Random Search, Genetic Algorithm, and Markov Chain Monte Carlo) to perform equivalent program transformation. Furthermore, we propose a deep reinforcement learning-based sequence generation (DRLSG) strategy to effectively guide the search process of generating clones that could escape from the detection. We then evaluate the ML-based detectors with the pairs of original and generated clones. We realize our method in a framework named CloneGen (stands for Clone Generator). CloneGen In evaluation, we challenge the three state-of-the-art ML-based detectors and four traditional detectors with the code clones after semantic-preserving transformations via the aid of CloneGen. Surprisingly, our experiments show that, despite the notable successes achieved by existing clone detectors, the ML models inside these detectors still cannot distinguish numerous clones produced by the code transformations in CloneGen. In addition, adversarial training of ML-based clone detectors using clones generated by CloneGen can improve their robustness and accuracy. Meanwhile, compared with the commonly-used heuristics, the DRLSG strategy has shown the best effectiveness in generating code clones to decrease the detection accuracy of the ML-based detectors. Our investigation reveals an explicable but always ignored robustness issue of the latest ML-based detectors. Therefore, we call for more attention to the robustness of these new ML-based detectors.

Deep learning frameworks play a key rule to bridge the gap between deep learning theory and practice. With the growing of safety- and security-critical applications built upon deep learning frameworks, their reliability is becoming increasingly important. To ensure the reliability of these frameworks, several efforts have been taken to study the causes and symptoms of bugs in deep learning frameworks, however, relatively little progress has been made in investigating the fault triggering conditions of those bugs. This paper presents the first comprehensive empirical study on fault triggering conditions in three widely-used deep learning frameworks (i.e., TensorFlow, MXNET and PaddlePaddle). We have collected 3,555 bug reports from GitHub repositories of these frameworks. A bug classification is performed based on fault triggering conditions, followed by the analysis of frequency distribution of different bug types and the evolution features. The correlations between bug types and fixing time are investigated. Moreover, we have also studied the root causes of Bohrbugs and Mandelbugs and investigated the important consequences of each bug type. Finally, the analysis of regression bugs in deep learning frameworks is conducted. We have revealed 12 important findings based on our empirical results and have provided 10 implications for developers and users.

Being able to automatically detect the performance issues in apps can significantly improve apps’ quality as well as having a positive influence on user satisfaction. Application Performance Management (APM) libraries are used to locate the apps’ performance bottleneck, monitor their behaviors at runtime, and identify potential security risks. Although app developers have been exploiting application performance management (APM) tools to capture these potential performance issues, most of them do not fully understand the internals of these APM tools and the effect on their apps. To fill this gap, in this paper, we conduct the first systematic study on APMs for apps by scrutinizing 25 widely-used APMs for Android apps and develop a framework named APMHunter for exploring the usage of APMs in Android apps. Using APMHunter, we conduct a large-scale empirical study on 500,000 Android apps to explore the usage patterns of APMs and discover the potential misuses of APMs. We obtain two major findings: 1) some APMs still employ deprecated permissions and approaches, which makes APMs fail to perform as expected; 2) inappropriate use of APMs can cause privacy leaks. Thus, our study suggests that both APM vendors and developers should design and use APMs scrupulously.

We present Supa, a value-flow-based demand-driven flow- and context-sensitive pointer analysis with strong updates for C and C++ programs. Supa enables computing points-to information via value-flow refinement, in environments with small time and memory budgets. We formulate Supa by solving a graph-reachability problem on an inter-procedural value-flow graph representing a program's def-use chains, which are pre-computed efficiently but over-approximately. To answer a client query (a request for a variable's points-to set), Supa reasons about the flow of values along the pre-computed def-use chains sparsely (rather than across all program points), by performing only the work necessary for the query (rather than analyzing the whole program). In particular, strong updates are performed to filter out spurious def-use chains through value-flow refinement as long as the total budget is not exhausted. We have implemented Supa on top of LLVM (4.0.0) together with a comprehensive micro-benchmark suite after a years-long effort (consisting of around 400 test cases, including hand-written ones and the ones extracted from real programs). We have evaluated Supa by choosing uninitialized pointer detection and C++ virtual table resolution as two major clients, using 24 real-world programs including 18 open-source C programs and 6 large CPU2000/2006 C++ benchmarks. For uninitialized pointer client, Supa achieves improved precision as the analysis budget increases, with its flow-sensitive (context-insensitive) analysis reaching 97.4 percent of that achieved by whole-program Sparse Flow-Sensitive analysis (SFS) by consuming about 0.18 seconds and 65 KB of memory per query, on average (with a budget of at most 10,000 value-flow edges per query). With context-sensitivity also considered, Supa becomes more precise for some programs but also incurs more analysis times. To further demonstrate the effectiveness of Supa, we have also evaluated Supa in resolving C++ virtual tables by querying the function pointers at every virtual callsite. Compared to analysis without strong updates for heap objects, Supa's demand-driven context-sensitive strong update analysis reduces 7.35 percent spurious virtual table targets with only 0.4 secs per query, on average.

Gas-based reduction roasting magnetic separation process was conducted to separate and recover iron and titanium from oxidized vanadium titano-magnetite (VTM). The effects of reduction temperature, gas composition, reduction time, grinding fineness, and magnetic field intensity on the efficiency of Fe-Ti separation were investigated. Under the condition of reduction temperature of 1050°C, reduction time of 120min and gas composition of P(H2)/P(H2+CO)=0.72, a magnetic product containing 84.5wt.% total iron with Fe recovery of 93.67% and a non-magnetic product containing 43.46wt.% TiO2with the TiO2recovery of 67.12% were obtained. The efficiency of Fe-Ti separation shows a rise first followed by a decline with the increases of P(H2)/P(H2+CO). Theoretical analyses, XRD, and SEM-EDS results demonstrate that H2has a higher reduction ability than CO on oxidized VTM while CO makes a better role in promoting the grown of metallic iron particles, which are both crucial for Fe-Ti separation.

C is the most widely used programming language for developing embedded software, operating systems, and device drivers. Unlike programs written in managed languages like Java, C programs rely on explicit memory management, and are therefore prone to memory leaks. Existing (static or dynamic) debugging tools only report leaks, but fixing them often requires considerable manual effort by inspecting a list of reported true and false alarms. How to develop on-demand lightweight techniques for automated leak fixing without introducing new memory errors remains challenging. In this paper, we introduce AutoFix, a fully automated leak-fixing approach for C programs by combining static and dynamic program analyses. Given a leaky allocation site reported by a static memory leak detector, AutoFix performs a graph reachability analysis to identify leaky paths on the value-flow slices of the program, and then conducts a liveness analysis to locate the program points for inserting fixes (i.e., the missing free calls) on the identified leaky paths. We have implemented AutoFix in LLVM-3.5.0 and evaluated it using five SPEC2000 benchmarks and three open-source applications. Experimental results show that AutoFix can safely fix all the memory leaks reported by a state-of-the-art static memory leak detector with small instrumentation overhead.

This paper presents a comprehensive survey on the development of Intel SGX (software guard extensions) processors and its applications. With the advent of SGX in 2013 and its subsequent development, the corresponding research works are also increasing rapidly. In order to get a more comprehensive literature review related to SGX, we have made a systematic analysis of the related papers in this area. We first search through five large-scale paper retrieval libraries by keywords (i.e., ACM Digital Library, IEEE/IET Electronic Library, SpringerLink, Web of Science, and Elsevier Science Direct). We read and analyze a total of 128 SGX-related papers. The first round of extensive study is conducted to classify them. The second round of intensive study is carried out to complete a comprehensive analysis of the paper from various aspects. We start with the working environment of SGX and make a conclusive summary of trusted execution environment (TEE). We then focus on the applications of SGX. We also review and study multifarious attack methods to SGX framework and some recent security improvements made on SGX. Finally, we summarize the advantages and disadvantages of SGX with some future research opportunities. We hope this review could help the existing and future research works on SGX and its application for both developers and users.