Yuan Xiao

Yuchen Chen, Yuan Xiao, Chunrong Fang et al.

The proliferation of large language models for code (CodeLMs) and open-source contributions has heightened concerns over unauthorized use of source code datasets. While watermarking provides a viable protection mechanism by embedding ownership signals, existing methods rely on detectable trigger-target patterns and are limited to source-code tasks, overlooking other scenarios such as decompilation tasks. In this paper, we propose DuCodeMark, a stealthy and robust dual-purpose watermarking method for code datasets that generalizes across both source-code tasks and decompilation tasks. DuCodeMark parses each code sample into an abstract syntax tree (AST), applies language-specific style transformations to construct stealthy trigger-target pairs, and injects repressible poisoned features into a subset of return-typed samples to enhance robustness against watermark removal or evasion. These features remain inactive during normal training but are activated upon watermark removal, degrading model performance. For verification, DuCodeMark employs a black-box method based on the independent-samples $t$-test. We conduct a comprehensive evaluation of DuCodeMark across 72 settings spanning two code tasks, two programming languages, three CodeLMs, and six decoding temperatures. The results demonstrate that it consistently achieves strong verifiability ($p < 0.05$), high stealthiness (suspicion rate $\leq$ 0.36), robustness against both watermark and poisoning attacks (recall $\leq$ 0.57), and a substantial drop in model performance upon watermark removal (Pass@1 drops by 28.6%), underscoring its practicality and resilience.

Yuan Xiao, Yuchen Chen, Shiqing Ma et al.

The robustness of neural network classifiers is important in the safety-critical domain and can be quantified by robustness verification. At present, efficient and scalable verification techniques are always sound but incomplete, and thus, the improvement of verified robustness results is the key criterion to evaluate the performance of incomplete verification approaches. The multi-variate function MaxPool is widely adopted yet challenging to verify. In this paper, we present Ti-Lin, a robustness verifier for MaxPool-based CNNs with Tight Linear Approximation. Following the sequel of minimizing the over-approximation zone of the non-linear function of CNNs, we are the first to propose the provably neuron-wise tightest linear bounds for the MaxPool function. By our proposed linear bounds, we can certify larger robustness results for CNNs. We evaluate the effectiveness of Ti-Lin on different verification frameworks with open-sourced benchmarks, including LeNet, PointNet, and networks trained on the MNIST, CIFAR-10, Tiny ImageNet and ModelNet40 datasets. Experimental results show that Ti-Lin significantly outperforms the state-of-the-art methods across all networks with up to 78.6% improvement in terms of the certified accuracy with almost the same time consumption as the fastest tool. Our code is available at https://github.com/xiaoyuanpigo/Ti-Lin-Hybrid-Lin.

Weisong Sun, Yuchen Chen, Chunrong Fang et al.

Neural code models (NCMs) have been widely used to address various code understanding tasks, such as defect detection. However, numerous recent studies reveal that such models are vulnerable to backdoor attacks. Backdoored NCMs function normally on normal/clean code snippets, but exhibit adversary-expected behavior on poisoned code snippets injected with the adversary-crafted trigger. It poses a significant security threat. Therefore, there is an urgent need for effective techniques to detect and eliminate backdoors stealthily implanted in NCMs. To address this issue, in this paper, we innovatively propose a backdoor elimination technique for secure code understanding, called EliBadCode. EliBadCode eliminates backdoors in NCMs by inverting/reverse-engineering and unlearning backdoor triggers. Specifically, EliBadCode first filters the model vocabulary for trigger tokens based on the naming conventions of specific programming languages to reduce the trigger search space and cost. Then, EliBadCode introduces a sample-specific trigger position identification method, which can reduce the interference of non-backdoor (adversarial) perturbations for subsequent trigger inversion, thereby producing effective inverted backdoor triggers efficiently. Backdoor triggers can be viewed as backdoor (adversarial) perturbations. Subsequently, EliBadCode employs a Greedy Coordinate Gradient algorithm to optimize the inverted trigger and designs a trigger anchoring method to purify the inverted trigger. Finally, EliBadCode eliminates backdoors through model unlearning. We evaluate the effectiveness of EliBadCode in eliminating backdoors implanted in multiple NCMs used for three safety-critical code understanding tasks. The results demonstrate that EliBadCode can effectively eliminate backdoors while having minimal adverse effects on the normal functionality of the model.

Tingxu Han, Wei Song, Weisong Sun et al.

With the development of deep learning, Neural Code Models (NCMs) such as CodeBERT and CodeLlama are widely used for code understanding tasks, including defect detection and code classification. However, recent studies have revealed that NCMs are vulnerable to adversarial examples, inputs with subtle perturbations that induce incorrect predictions while remaining difficult to detect. Existing defenses address this issue via data augmentation to empirically improve robustness, but they are costly, offer no theoretical robustness guarantees, and typically require white-box access to model internals, such as gradients. To address the above challenges, we propose ENBECOME, a novel black-box training-free and lightweight adversarial defense. ENBECOME is designed to both enhance empirical robustness and report certified robustness boundaries for NCMs. ENBECOME operates solely during inference, introducing random, semantics-preserving perturbations to input code snippets to smooth the NCM's decision boundaries. This smoothing enables ENBECOME to formally certify a robustness radius within which adversarial examples can never induce misclassification, a property known as certified robustness. We conduct comprehensive experiments across multiple NCM architectures and tasks. Results show that ENBECOME significantly reduces attack success rates while maintaining high accuracy. For example, in defect detection, it reduces the average ASR from 42.43% to 9.74% with only a 0.29% drop in accuracy. Results show that ENBECOME significantly reduces attack success rates while maintaining high accuracy. For example, in defect detection, it reduces the average ASR from 42.43% to 9.74% with only a 0.29% drop in accuracy. Furthermore, ENBECOME achieves an average certified robustness radius of 1.63, meaning that adversarial modifications to no more than 1.63 identifiers are provably ineffective.

Yuan Xiao, Jiaming Wang, Yuchen Chen et al.

The widespread availability of large-scale code datasets has accelerated the development of code large language models (CodeLLMs), raising concerns about unauthorized dataset usage. Dataset poisoning offers a proactive defense by reducing the utility of such unauthorized training. However, existing poisoning methods often require full dataset poisoning and introduce transformations that break code compilability. In this paper, we introduce FunPoison, a functionality-preserving poisoning approach that injects short, compilable weak-use fragments into executed code paths. FunPoison leverages reusable statement-level templates with automatic repair and conservative safety checking to ensure side-effect freedom, while a type-aware synthesis module suppresses static analysis warnings and enhances stealth. Extensive experiments show that FunPoison achieves effective poisoning by contaminating only 10% of the dataset, while maintaining 100% compilability and functional correctness, and remains robust against various advanced code sanitization techniques.

Yuan Xiao, Shiqing Ma, Juan Zhai et al.

The robustness of convolutional neural networks (CNNs) is vital to modern AI-driven systems. It can be quantified by formal verification by providing a certified lower bound, within which any perturbation does not alter the original input's classification result. It is challenging due to nonlinear components, such as MaxPool. At present, many verification methods are sound but risk losing some precision to enhance efficiency and scalability, and thus, a certified lower bound is a crucial criterion for evaluating the performance of verification tools. In this paper, we present MaxLin, a robustness verifier for MaxPool-based CNNs with tight linear approximation. By tightening the linear approximation of the MaxPool function, we can certify larger certified lower bounds of CNNs. We evaluate MaxLin with open-sourced benchmarks, including LeNet and networks trained on the MNIST, CIFAR-10, and Tiny ImageNet datasets. The results show that MaxLin outperforms state-of-the-art tools with up to 110.60% improvement regarding the certified lower bound and 5.13 $\times$ speedup for the same neural networks. Our code is available at https://github.com/xiaoyuanpigo/maxlin.

Yuan Xiao, Mengyuan Li, Sanchuan Chen et al.

Intel Software Guard Extension (SGX) offers software applications enclave to protect their confidentiality and integrity from malicious operating systems. The SSL/TLS protocol, which is the de facto standard for protecting transport-layer network communications, has been broadly deployed for a secure communication channel. However, in this paper, we show that the marriage between SGX and SSL may not be smooth sailing. Particularly, we consider a category of side-channel attacks against SSL/TLS implementations in secure enclaves, which we call the control-flow inference attacks. In these attacks, the malicious operating system kernel may perform a powerful man-in-the-kernel attack to collect execution traces of the enclave programs at page, cacheline, or branch level, while positioning itself in the middle of the two communicating parties. At the center of our work is a differential analysis framework, dubbed Stacco, to dynamically analyze the SSL/TLS implementations and detect vulnerabilities that can be exploited as decryption oracles. Surprisingly, we found exploitable vulnerabilities in the latest versions of all the SSL/TLS libraries we have examined. To validate the detected vulnerabilities, we developed a man-in-the-kernel adversary to demonstrate Bleichenbacher attacks against the latest OpenSSL library running in the SGX enclave (with the help of Graphene) and completely broke the PreMasterSecret encrypted by a 4096-bit RSA public key with only 57286 queries. We also conducted CBC padding oracle attacks against the latest GnuTLS running in Graphene-SGX and an open-source SGX-implementation of mbedTLS (i.e., mbedTLS-SGX) that runs directly inside the enclave, and showed that it only needs 48388 and 25717 queries, respectively, to break one block of AES ciphertext. Empirical evaluation suggests these man-in-the-kernel attacks can be completed within 1 or 2 hours.

Haocheng Huang, Yuchen Chen, Weisong Sun et al.

Constructing and curating high-quality code datasets requires significant resources, making them valuable intellectual property. Unfortunately, these datasets currently face severe risks of unauthorized use. Although digital watermarking offers a post hoc mechanism for copyright authentication, existing methods are predominantly based on the co-occurrence pattern, which is not robust and is susceptible to watermark detection and removal attacks. In this paper, we propose PuzzleMark, a robust watermarking method for code datasets. To reduce the risk of watermark exposure, PuzzleMark introduces a carrier selection strategy that leverages code complexity to evaluate the suitability of code snippets as watermark carriers, and selects those with high suitability for watermarking. To enhance the robustness of the watermark, PuzzleMark proposes a novel concatenation pattern to replace the traditional co-occurrence pattern, and implements two watermarking strategies through variable name concatenation. PuzzleMark adaptively embeds watermarks based on the inherent characteristics of the code, making it more stealthy while maintaining design simplicity. For watermark verification, PuzzleMark employs Fisher's exact test to verify suspicious models under a black-box setting. Experimental results demonstrate that PuzzleMark achieves a 100% verification success rate and a 0% false positive rate, with negligible impact on model performance. Both our human study and our evaluation using four state-of-the-art watermark detection methods show that PuzzleMark exhibits strong imperceptibility, with an average suspicious rate $\leq$ 0.24 and an average recall $\leq$ 30.41%, respectively. As a practical digital watermarking method, PuzzleMark provides strong protection for the intellectual property of code datasets and offers new insights for future research.

Abdullah-Al-Zubaer Imran, Chao Huang, Hui Tang et al.

Semi-supervised learning has recently been attracting attention as an alternative to fully supervised models that require large pools of labeled data. Moreover, optimizing a model for multiple tasks can provide better generalizability than single-task learning. Leveraging self-supervision and adversarial training, we propose a novel general purpose semi-supervised, multiple-task model---namely, self-supervised, semi-supervised, multitask learning (S$^4$MTL)---for accomplishing two important tasks in medical imaging, segmentation and diagnostic classification. Experimental results on chest and spine X-ray datasets suggest that our S$^4$MTL model significantly outperforms semi-supervised single task, semi/fully-supervised multitask, and fully-supervised single task models, even with a 50\% reduction of class and segmentation labels. We hypothesize that our proposed model can be effective in tackling limited annotation problems for joint training, not only in medical imaging domains, but also for general-purpose vision tasks.

Yuan Xiao, Yinqian Zhang, Radu Teodorescu

SPEculative Execution side Channel Hardware (SPEECH) Vulnerabilities have enabled the notorious Meltdown, Spectre, and L1 terminal fault (L1TF) attacks. While a number of studies have reported different variants of SPEECH vulnerabilities, they are still not well understood. This is primarily due to the lack of information about microprocessor implementation details that impact the timing and order of various micro-architectural events. Moreover, to date, there is no systematic approach to quantitatively measure SPEECH vulnerabilities on commodity processors. This paper introduces SPEECHMINER, a software framework for exploring and measuring SPEECH vulnerabilities in an automated manner. SPEECHMINER empirically establishes the link between a novel two-phase fault handling model and the exploitability and speculation windows of SPEECH vulnerabilities. It enables testing of a comprehensive list of exception-triggering instructions under the same software framework, which leverages covert-channel techniques and differential tests to gain visibility into the micro-architectural state changes. We evaluated SPEECHMINER on 9 different processor types, examined 21 potential vulnerability variants, confirmed various known attacks, and identified several new variants.

Ping Lu, Yanyan Zhang, Jianxiong Chen et al.

We propose a predictive neural network architecture that can be utilized to update reference velocity models as inputs to the full waveform inversion. Deep learning models are explored to augment velocity model building workflows during processing the 3D seismic volume in salt-prone environments. Specifically, a neural network architecture, with 3D convolutional, de-convolutional layers, and 3D max-pooling, is designed to take standard amplitude 3D seismic volumes as an input. Enhanced data augmentations through generative adversarial networks and a weighted loss function enable the network to train with few sparsely annotated slices. Batch normalization is also applied for faster convergence. A 3D probability cube for salt bodies and inclusions is generated through ensembles of predictions from multiple models in order to reduce variance. Velocity models inferred from the proposed networks provide opportunities for FWI forward models to converge faster with an initial condition closer to the true model. In addition, in each iteration step, the probability cubes of salt bodies and inclusions inferred from the proposed networks can be used as a regularization term within the FWI forward modelling, which may result in an improved velocity model estimation while the output of seismic migration can be utilized as an input of the 3D neural network for subsequent iterations.

Pei Yan, Yihua Tan, Yuan Xiao et al.

This paper presents an entirely unsupervised interest point training framework by jointly learning detector and descriptor, which takes an image as input and outputs a probability and a description for every image point. The objective of the training framework is formulated as joint probability distribution of the properties of the extracted points. The essential properties are selected as sparsity, repeatability and discriminability which are formulated by the probabilities. To maximize the objective efficiently, latent variable is introduced to represent the probability of that a point satisfies the required properties. Therefore, original maximization can be optimized with Expectation Maximization algorithm (EM). Considering high computation cost of EM on large scale image set, we implement the optimization process with an efficient strategy as Mini-Batch approximation of EM (MBEM). In the experiments both detector and descriptor are instantiated with fully convolutional network which is named as Property Network (PN). The experiments demonstrate that PN outperforms state-of-the-art methods on a number of image matching benchmarks without need of retraining. PN also reveals that the proposed training framework has high flexibility to adapt to diverse types of scenes.

Guoxing Chen, Sanchuan Chen, Yuan Xiao et al.

This paper presents SgxPectre Attacks that exploit the recently disclosed CPU bugs to subvert the confidentiality and integrity of SGX enclaves. Particularly, we show that when branch prediction of the enclave code can be influenced by programs outside the enclave, the control flow of the enclave program can be temporarily altered to execute instructions that lead to observable cache-state changes. An adversary observing such changes can learn secrets inside the enclave memory or its internal registers, thus completely defeating the confidentiality guarantee offered by SGX. To demonstrate the practicality of our SgxPectre Attacks, we have systematically explored the possible attack vectors of branch target injection, approaches to win the race condition during enclave's speculative execution, and techniques to automatically search for code patterns required for launching the attacks. Our study suggests that any enclave program could be vulnerable to SgxPectre Attacks since the desired code patterns are available in most SGX runtimes (e.g., Intel SGX SDK, Rust-SGX, and Graphene-SGX). Most importantly, we have applied SgxPectre Attacks to steal seal keys and attestation keys from Intel signed quoting enclaves. The seal key can be used to decrypt sealed storage outside the enclaves and forge valid sealed data; the attestation key can be used to forge attestation signatures. For these reasons, SgxPectre Attacks practically defeat SGX's security protection. This paper also systematically evaluates Intel's existing countermeasures against SgxPectre Attacks and discusses the security implications.