RENN: Efficient Reverse Execution with Neural-Network-Assisted Alias Analysis

Mu, Dongliang; Guo, Wenbo; Cuevas, Alejandro; Chen, Yueqi; Gai, Jinxuan; Xing, Xinyu; Mao, Bing; Song, Chengyu

doi:10.1109/ase.2019.00090

Cited by 13 publications

(13 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unlike approaches mentioned above, recent works [16], [17] suggest approaches that aim at specific goals rather than representing whole program behavior. These works use dynamic information of data (or state) to recognize that this information plays an important role in program behavior modeling.…”

Section: Modeling For Specific Purposementioning

confidence: 99%

Data Embedding Scheme for Efficient Program Behavior Modeling With Neural Networks

Ahn

Bae

et al. 2022

IEEE Trans. Emerg. Top. Comput. Intell.

View full text Add to dashboard Cite

As modern programs grow in size and complexity, the importance of program behavior modeling is emerging in various areas. Because of the large amount of data generated by a target program and the difficulty of runtime analysis, previous works in these areas employ deep learning. However, they did not sufficiently consider the input of a target program, since, in our view, program behavior is a history of computational steps consisting of a function and its input arguments. A naive, intuitive way to embed the value of x as it is in a vector representation creates a tremendously large vector size. Instead, we found that all the values inducing the same runtime behavior can be represented as one identical characteristic value (CV). In this paper, we show that not only can a characteristic value sequence replace the argument input, but it is also efficient to use it as an input vector for a neural network. This efficiency comes from modeling the whole program with multiple LSTM-RNN models and reducing the input space of the neural network. To demonstrate the effectiveness of this replacement, we performed experiments on the problem of program behavior anomaly detection. Our results show that our model achieves better detection performance compared to previous models and similar detection performance even with smaller model sizes. We also provide a visualization of the embedded vectors extracted from the embedding layer in the neural network model to prove that the CV sequence well represents the arguments.

show abstract

Section: Modeling For Specific Purposementioning

confidence: 99%

Data Embedding Scheme for Efficient Program Behavior Modeling With Neural Networks

Ahn

Bae

et al. 2022

IEEE Trans. Emerg. Top. Comput. Intell.

View full text Add to dashboard Cite

show abstract

“…For example, EKLAVYA [12] learns function type signatures. DeepVSA [25] and RENN [38] learn memory alias dependencies. DeepBinDiff [22] and GEMINI [62] learn function similarity by neural embeddings.…”

Section: Related Workmentioning

confidence: 99%

XDA: Accurate, Robust Disassembly with Transfer Learning

Pei

Guan

Williams-King

et al. 2021

Proceedings 2021 Network and Distributed System Security Symposium

View full text Add to dashboard Cite

Accurate and robust disassembly of stripped binaries is challenging. The root of the difficulty is that highlevel structures, such as instruction and function boundaries, are absent in stripped binaries and must be recovered based on incomplete information. Current disassembly approaches rely on heuristics or simple pattern matching to approximate the recovery, but these methods are often inaccurate and brittle, especially across different compiler optimizations. We present XDA, a transfer-learning-based disassembly framework that learns different contextual dependencies present in machine code and transfers this knowledge for accurate and robust disassembly. We design a self-supervised learning task motivated by masked Language Modeling to learn interactions among byte sequences in binaries. The outputs from this task are byte embeddings that encode sophisticated contextual dependencies between input binaries' byte tokens, which can then be finetuned for downstream disassembly tasks. We evaluate XDA's performance on two disassembly tasks, recovering function boundaries and assembly instructions, on a collection of 3,121 binaries taken from SPEC CPU2017, SPEC CPU2006, and the BAP corpus. The binaries are compiled by GCC, ICC, and MSVC on x86/x64 Windows and Linux platforms over 4 optimization levels. XDA achieves 99.0% and 99.7% F1 score at recovering function boundaries and instructions, respectively, surpassing the previous state-of-the-art on both tasks. It also maintains speed on par with the fastest MLbased approach and is up to 38× faster than handwritten disassemblers like IDA Pro. We release the code of XDA at https://github.com/CUMLSec/XDA. Previous research has explored using Machine Learning (ML) to address these challenges. The resulting models surpassed the accuracy of traditional disassemblers at both recovering assembly instructions [58] and function boundaries [7], [48], and are easier to maintain. However, current ML-based approaches still face two critical challenges: Accuracy. Recent research has shown that previous ML-based methods are likely not as accurate as reported [5]. This is because their accuracies were inadvertently measured on testing data with significant overlap with training data, and therefore their performance may not generalize to real-world binaries. More concretely, Andriesse et al. [5] showed that the F1 score of ByteWeight [7] degrades from the reported 97% to 65% when evaluated on a dataset without overlap. Robustness. The current most accurate ML approach for recovering function boundaries, bidirectional Recurrent Neural

show abstract

“…Code search [15,44,45,134,138], Code classification [30,74,154] code readability classification/prediction [105,117], Function type inferring [53,103] Code generation [35,115], Code Summarization [75,135], Code Decompilation [66,71] Code change generation [130], Data structure classification [108], Reverse execution [111] Design pattern detection [127], Technical debt detection [118], Story points prediction [21] Inconsistent method name refactoring [92], Stable patch detection [55] Defect Defect prediction [94,98,129,136,137,140], Vulnerability prediction [27,38,52,151] 8 24 Bug localization/detection [59,72,81,139,144,155] , Code repair [10,132,141] Code smell detection …”

Section: Codementioning

confidence: 99%

On the Replicability and Reproducibility of Deep Learning in Software Engineering

Liu,

Gao,

Xia

et al. 2020

Preprint

View full text Add to dashboard Cite

Deep learning (DL) techniques have gained significant popularity among software engineering (SE) researchers in recent years. This is because they can often solve many SE challenges without enormous manual feature engineering effort and complex domain knowledge. Although many DL studies have reported substantial advantages over other state-of-the-art models on effectiveness, they often ignore two factors: (1) replicability -whether the reported experimental result can be approximately reproduced in high probability with the same DL model and the same data; and (2) reproducibility -whether one reported experimental findings can be reproduced by new experiments with the same experimental protocol and DL model, but different sampled real-world data. Unlike traditional machine learning (ML) models, DL studies commonly overlook these two factors and declare them as minor threats or leave them for future work. This is mainly due to high model complexity with many manually set parameters and the time-consuming optimization process. In this study, we conducted a literature review on 93 DL studies recently published in twenty SE journals or conferences. Our statistics show the urgency of investigating these two factors in SE, where only 10.8% of the studies discussed any research questions affecting replicability and/or reproducibility. More than 74.2% of the studies do not even share source code and data to support the replicability of their complex models. Moreover, we re-ran four representative DL models in SE. Experimental results show the importance of replicability and reproducibility, where the reported performance of a DL model could not be replicated for an unstable optimization process. Reproducibility could be substantially compromised if the model training is not convergent, or if performance is sensitive to the size of vocabulary and testing data. It is therefore urgent for the SE community to provide a long-lasting link to a replication package, enhance DL-based solution stability and convergence, and avoid performance sensitivity on different sampled data.CCS Concepts: • Software and its engineering → Software maintenance tools.

show abstract

RENN: Efficient Reverse Execution with Neural-Network-Assisted Alias Analysis

Cited by 13 publications

References 25 publications

Data Embedding Scheme for Efficient Program Behavior Modeling With Neural Networks

Data Embedding Scheme for Efficient Program Behavior Modeling With Neural Networks

XDA: Accurate, Robust Disassembly with Transfer Learning

On the Replicability and Reproducibility of Deep Learning in Software Engineering

Contact Info

Product

Resources

About