Investigating the spatiotemporal dynamics of the human brain is a complex challenge due to the intricate nature of brain networks, and the limitations of current analytical methods. Herein, we introduce the 4D functional Magnetic Resonance Imaging (fMRI) CrossFormer (4DfCF), a novel vision transformer architecture designed to process high-dimensional 4D fMRI data. This model integrates temporal and spatial dimensions to effectively learn and predict cognitive and clinical outcomes. We further evaluated the 4DfCF on three benchmark datasets: Attention Deficit Hyperactivity Disorder-200 (ADHD-200), Alzheimer's Disease Neuroimaging Initiative (ADNI), and Autism Brain Imaging Data Exchange (ABIDE). The results showed that our model consistently outperforms state-of-the-art baseline models, achieving an accuracy improvement of 5-10%, a precision increase of 4-8%, a recall enhancement of 6-9%, and an F1-score boost of 7-11% . Additionally, the 4D fMRI CrossFormer-Tiny variant demonstrated greater efficiency than existing methods, using 20% fewer computational resources and achieving 30% faster training times. Pre-training experiments further reveal that models pre-trained on one dataset and fine-tuned on another achieved faster convergence and higher accuracy, with the Autism Brain Imaging Data Exchange (ABIDE) pre-trained models showing the best performance. Additionally, we employed an explainable AI method to identify the brain regions associated with disease diagnosis. Overall, our findings highlight the potential of the 4DfCF to advance precision neuroscience through efficient and scalable analysis of complex fMRI data.

Deep neural network (DNN) models are susceptible to adversarial samples in white-box and opaque environments. Although previous studies have shown high attack success rates, coupling DNN models with interpretation models could offer a sense of security when a human expert is involved. However, in white-box environments, interpretable deep learning systems (IDLSes) have been shown to be vulnerable to malicious manipulations. As access to the components of IDLSes is limited in opaque settings, it becomes more challenging for the adversary to fool the system. In this work, we propose aQuery-efficientScore-based opaque attack against IDLSes, which requires no knowledge of the target model and its coupled interpretation model. By continuously refining the adversarial samples created based on feedback scores from the IDLS, our approach effectively reduces the number of model queries and navigates the search space to identify perturbations that can fool the system. We evaluate the attack's effectiveness on four convolutional neural network (CNN) models and two interpretation models, using both ImageNet and CIFAR datasets. Our results show that the proposed approach is query-efficient with a high attack success rate that can reach more than 95%, and an average transferability success rate of 69%. We have also demonstrated that our attack is resilient against various preprocessing defense techniques.

Deep learning methods have gained increasing attention in various applications due to their outstanding performance. For exploring how this high performance relates to the proper use of data artifacts and the accurate problem formulation of a given task, interpretation models have become a crucial component in developing deep learning-based systems. Interpretation models enable the understanding of the inner workings of deep learning models and offer a sense of security in detecting the misuse of artifacts in the input data. Similar to prediction models, interpretation models are also susceptible to adversarial inputs. This work introduces two attacks, AdvEdge and AdvEdge$^{+}$, which deceive both the target deep learning model and the coupled interpretation model. We assess the effectiveness of proposed attacks against four deep learning model architectures coupled with four interpretation models that represent different categories of interpretation models. Our experiments include the implementation of attacks using various attack frameworks. We also explore the attack resilience against three general defense mechanisms and potential countermeasures. Our analysis shows the effectiveness of our attacks in terms of deceiving the deep learning models and their interpreters, and highlights insights to improve and circumvent the attacks.

Establishing trust and helping experts debug and understand the inner workings of deep learning models, interpretation methods are increasingly coupled with these models, building interpretable deep learning systems. However, adversarial attacks pose a significant threat to public trust by making interpretations of deep learning models confusing and difficult to understand. In this paper, we present a novel Single-class target-specific ADVersarial attack called SingleADV. The goal of SingleADV is to generate a universal perturbation that deceives the target model into confusing a specific category of objects with a target category while ensuring highly relevant and accurate interpretations. The universal perturbation is stochastically and iteratively optimized by minimizing the adversarial loss that is designed to consider both the classifier and interpreter costs in targeted and non-targeted categories. In this optimization framework, ruled by the first- and second-moment estimations, the desired loss surface promotes high confidence and interpretation scores of adversarial samples. By avoiding unintended misclassification of samples from other categories, SingleADV enables more effective targeted attacks on interpretable deep learning systems in both white-box and black-box scenarios. To evaluate the effectiveness of SingleADV, we conduct experiments using four different model architectures (ResNet-50, VGG-16, DenseNet-169, and Inception-V3) coupled with three interpretation models (CAM, Grad, and MASK). Through extensive empirical evaluation, we demonstrate that SingleADV effectively deceives target deep learning models and their associated interpreters under various conditions and settings. Our results show that the performance of SingleADV is effective, with an average attack success rate of 74% and prediction confidence exceeding 77% on successful adversarial samples. Furthermore, we discuss several countermeasures against SingleADV, including a transfer-based learning approach and existing preprocessing defenses.

Despite the benefits of the Internet of Things (IoT), the growing influx of IoT-specific malware coordinating large-scale cyberattacks via infected IoT devices has created a substantial threat to the Internet ecosystem. Assessing IoT systems’ security and developing mitigation measures to prevent the spread of IoT malware is therefore critical. Furthermore, for training and testing the fidelity of cyber security-based Machine Learning (ML) and Deep Learning (DL) approaches, the collection and exploration of information from multiple sources from the IoT are crucial. In this regard, we propose a multitask DL model for detecting IoT malware. Our proposed Long Short-Term Memory (LSTM) based model efficiently performs two tasks: 1) determination of whether the provided traffic is benign or malicious, and 2) determination of the malware type for identifying malicious network traffic. We used large-scale traffic data of 145.pcapfiles of benign and malicious traffic collected from 18 different IoT devices. We performed a time-series analysis on the packets of traffic flows, which were then used to train the proposed model. The features extracted from the dataset were categorized into three modalities: flow-related, traffic flag-related, and packet payload-related features. A feature selection approach was employed at the feature and modality levels, and the best modalities and features were utilized for performance enhancement. For tasks 1 and 2 and multitask classification, the flow-related and flag-related modalities showed the best testing accuracies of 92.63%, 88.45%, and 95.83%, respectively.

Deep neural network models are susceptible to malicious manipulations even in the black-box settings. Providing explanations for DNN models offers a sense of security by human involvement, which reveals whether the sample is benign or adversarial even though previous studies achieved a high attack success rate. However, interpretable deep learning systems (IDLSes) are shown to be susceptible to adversarial manipulations in white-box settings. Attacking IDLSes in black-box settings is challenging and remains an open research domain. In this work, we propose a black-box version of the white-box AdvEdge approach against IDLSes, which is query-efficient and gradient-free without obtaining any knowledge of the target DNN model and its coupled interpreter. Our approach takes advantage of transfer-based and score-based techniques using the effective microbial genetic algorithm (MGA). We achieve a high attack success rate with a small number of queries and high similarity in interpretations between adversarial and benign samples.

Deep Learning is rapidly evolving to the point that it can be used in crucial safety and security applications, including self-driving vehicles, surveillance, drones, and robots. However, these deep learning models are vulnerable to attacks based on adversarial samples that are undetectable to the human eye but cause the model to misbehave. There is an increasing demand for comprehensive and in-depth analysis of behaviors of various attacks and the possible defenses against common deep learning models under several adversarial scenarios. In this study, we conducted four separate investigations. First, we examine the relationship between the model's complexity and its robustness against the studied attacks. Second, the connection between the performance and diversity of models is examined. Third, the first and second experiments were tested across different datasets to explore the impact of the dataset on the performance of the model. Four, throughout the defense strategies, the model behavior is extensively investigated. The code, trained models, and detailed settings and results are available at: https://github.com/InfoLab-SKKU/ML-Adversarial-Attacks-Analysis

The rapid pace of malware development and the widespread use of code obfuscation, polymorphism, and morphing techniques pose a considerable challenge to detecting and analyzing malware. Today, it is difficult for antivirus applications to use traditional signature-based detection methods to detect morphing malware. Thus, the emergence of structure graph-based detection methods has become a hope to solve this challenge. In this work, we propose a method for detecting malware using graphs' spectral heat and wave signatures, which are efficient and size- and permutation-invariant. We extracted 250 and 1,000 heat and wave representations, and we trained and tested heat and wave representations on eight machine learning classifiers. We used a dataset of 37,537 unpacked Windows malware executables and extracted the control flow graph (CFG) of each windows malware to obtain the spectral representations. Our experimental results showed that by using heat and wave spectral graph theory, the best malware analysis accuracy reached 95.9%.

In a hospital, accurate and rapid mortality prediction of Length of Stay (LOS) is essential since it is one of the essential measures in treating patients with severe diseases. When predictions of patient mortality and readmission are combined, these models gain a new level of significance. Therefore, the most expensive components of patient care are LOS and readmission rates. Several studies have assessed readmission to the hospital as a single-task issue. The performance, robustness, and stability of the model increase when many correlated tasks are optimized. This study develops multimodal multitasking Long Short-Term Memory (LSTM) Deep Learning (DL) model that can predict both LOS and readmission for patients using multi-sensory data from 47 patients. Continuous sensory data is divided into eight sections, each of which is recorded for an hour. The time steps are constructed using a dual 10-second window-based technique, resulting in six steps per hour. The 30 statistical features are computed by transforming the sensory input into the resulting vector. The proposed multitasking model predicts 30-day readmission as a binary classification problem and LOS as a regression task by constructing discrete time-step data based on the length of physical activity during a hospital stay. The proposed model is compared to a random forest for a single-task problem (classification or regression) because typical machine learning algorithms are unable to handle the multitasking challenge. In addition, sensory data combined with other cost-effective modalities such as demographics, laboratory tests, and comorbidities to construct reliable models for personalized, cost-effective, and medically acceptable prediction. With a high accuracy of 94.84%, the proposed multitask multimodal DL model classifies the patient's readmission status and determines the patient's LOS in hospital with a minimal Mean Square Error (MSE) of 0.025 and Root Mean Square Error (RMSE) of 0.077, which is promising, effective, and trustworthy.

Successful software authorship de-anonymization has both software forensics applications and privacy implications. However, the process requires an efficient extraction of authorship attributes. The extraction of such attributes is very challenging, due to various software code formats from executable binaries with different toolchain provenance to source code with different programming languages. Moreover, the quality of attributes is bounded by the availability of software samples to a certain number of samples per author and a specific size for software samples. To this end, this work proposes a deep Learning-based approach for software authorship attribution, that facilitates large-scale, format-independent, language-oblivious, and obfuscation-resilient software authorship identification. This proposed approach incorporates the process of learning deep authorship attribution using a recurrent neural network, and ensemble random forest classifier for scalability to de-anonymize programmers. Comprehensive experiments are conducted to evaluate the proposed approach over the entire Google Code Jam (GCJ) dataset across all years (from 2008 to 2016) and over real-world code samples from 1,987 public repositories on GitHub. The results of our work show high accuracy despite requiring a smaller number of samples per author. Experimenting with source-code, our approach allows us to identify 8,903 GCJ authors, the largest-scale dataset used by far, with an accuracy of 92.3%. Using the real-world dataset, we achieved an identification accuracy of 94.38% for 745 C programmers on GitHub. Moreover, the proposed approach is resilient to language-specifics, and thus it can identify authors of four programming languages (e.g., C, C++, Java, and Python), and authors writing in mixed languages (e.g., Java/C++, Python/C++). Finally, our system is resistant to sophisticated obfuscation (e.g., using C Tigress) with an accuracy of 93.42% for a set of 120 authors. Experimenting with executable binaries, our approach achieves 95.74% for identifying 1,500 programmers of software binaries. Similar results were obtained when software binaries are generated with different compilation options, optimization levels, and removing of symbol information. Moreover, our approach achieves 93.86% for identifying 1,500 programmers of obfuscated binaries using all features adopted in Obfuscator-LLVM tool.

Smartphones have become crucial for our daily life activities and are increasingly loaded with our personal information to perform several sensitive tasks, including, mobile banking and communication, and are used for storing private photos and files. Therefore, there is a high demand for applying usable authentication techniques that prevent unauthorized access to sensitive information. In this article, we propose AUToSen, a deep-learning-based active authentication approach that exploits sensors in consumer-grade smartphones to authenticate a user. Unlike conventional approaches, AUToSen is based on deep learning to identify user distinct behavior from the embedded sensors with and without the user's interaction with the smartphone. We investigate different deep learning architectures in modeling and capturing users' behavioral patterns for the purpose of authentication. Moreover, we explore the sufficiency of sensory data required to accurately authenticate users. We evaluate AUToSen on a real-world data set that includes sensors data of 84 participants' smartphones collected using our designed data-collection application. The experiments show that AUToSen operates accurately using readings of only three sensors (accelerometer, gyroscope, and magnetometer) with a high authentication frequency, e.g., one authentication attempt every 0.5 s. Using sensory data of one second enables an authentication F1-score of approximately 98%, false acceptance rate (FAR) of 0.95%, false rejection rate (FRR) of 6.67%, and equal error rate (EER) of 0.41%. While using sensory data of half a second enables an authentication F1-score of 97.52%, FAR of 0.96%, FRR of 8.08%, and EER of 0.09%. Moreover, we investigate the effects of using different sensory data at variable sampling periods on the performance of the authentication models under various settings and learning architectures.

Abstract Most authorship identification schemes assume that code samples are written by a single author. However, real software projects are typically the result of a team effort, making it essential to consider a finegrained multi-author identification in a single code sample, which we address with Multi-χ. Multi-χ leverages a deep learning-based approach for multi-author identification in source code, is lightweight, uses a compact representation for efficiency, and does not require any code parsing, syntax tree extraction, nor feature selection. In Multi-χ, code samples are divided into small segments, which are then represented as a sequence ofn-dimensional term representations. The sequence is fed into an RNN-based verification model to assist a segment integration process which integrates positively verified segments, i.e., integrates segments that have a high probability of being written by one author. Finally, the resulting segments from the integration process are represented using word2vec or TF-IDF and fed into the identification model. We evaluate Multi-χ with several Github projects (Caffe, Facebook’s Folly, Tensor-Flow, etc.) and show remarkable accuracy. For example, Multi-χ achieves an authorship example-based accuracy (A-EBA) of 86.41% and per-segment authorship identification of 93.18% for identifying 562 programmers. We examine the performance against multiple dimensions and design choices, and demonstrate its effectiveness.

Efficient extraction of code authorship attributes is key for successful identification. However, the extraction of such attributes is very challenging, due to various programming language specifics, the limited number of available code samples per author, and the average code lines per file, among others. To this end, this work proposes a Deep Learning-based Code Authorship Identification System (DL-CAIS) for code authorship attribution that facilitates large-scale, language-oblivious, and obfuscation-resilient code authorship identification. The deep learning architecture adopted in this work includes TF-IDF-based deep representation using multiple Recurrent Neural Network (RNN) layers and fully-connected layers dedicated to authorship attribution learning. The deep representation then feeds into a random forest classifier for scalability to de-anonymize the author. Comprehensive experiments are conducted to evaluate DL-CAIS over the entire Google Code Jam (GCJ) dataset across all years (from 2008 to 2016) and over real-world code samples from 1987 public repositories on GitHub. The results of our work show the high accuracy despite requiring a smaller number of files per author. Namely, we achieve an accuracy of 96% when experimenting with 1,600 authors for GCJ, and 94.38% for the real-world dataset for 745 C programmers. Our system also allows us to identify 8,903 authors, the largest-scale dataset used by far, with an accuracy of 92.3%. Moreover, our technique is resilient to language-specifics, and thus it can identify authors of four programming languages (e.g. C, C++, Java, and Python), and authors writing in mixed languages (e.g. Java/C++, Python/C++). Finally, our system is resistant to sophisticated obfuscation (e.g. using C Tigress) with an accuracy of 93.42% for a set of 120 authors.

Many techniques have been proposed for reducing errors during text input on touchscreens. However, the majority of these techniques suffer from the same limitation, i.e., the keyboard keys are overcrowded on a small screen, resulting in high error rates and slow text inputs. To address this situation and resolve the problems associated with overcrowdedness, we introduce a new text-entry method called the “UOIT keyboard.” The idea behind the UOIT keyboard is to compose letters using “drawing-like typing” on the UOIT keyboard, which has 13 large keys that replace the 26 small keys that exist in the QWERTY keyboard. We describe the design, keys, and mechanism of the UOIT keyboard. A 24-participant user study was conducted to evaluate the speed and accuracy of the proposed entry method as compared with the QWERTY and multitap entry methods. As part of the evaluation, a questionnaire was used to collect participants' preferences. The UOIT keyboard has a mean entry speed of 11.3 words/min. The UOIT keyboard significantly reduces the typing errors with 3.8% total error rate comparing with 11.2% and 16.3% for QWERTY and multitap entry methods, respectively.

Connectivity and trust within social networks have been exploited to build applications on top of these networks, including information dissemination, Sybil defenses, and anonymous communication systems. In these networks, and for such applications, connectivity ensures good performance of applications while trust is assumed to always hold, so as collaboration and good behavior are always guaranteed. In this paper, we study the impact of differential behavior of users on performance in typical social network-based information dissemination applications. We classify users into either collaborative or rational (probabilistically collaborative) and study the impact of this classification and the associated behavior of users on the performance on such applications. By experimenting with real-world social network traces, we make several interesting observations. First, we show that some of the existing social graphs have high routing costs, demonstrating poor structure that prevents their use in such applications. Second, we study the factors that make probabilistically collaborative nodes important for the performance of the routing protocol within the entire network and demonstrate that the importance of these nodes stems from their topological features rather than their percentage of all the nodes within the network.

Sensor nodes are usually vulnerable to be compromised due to their unattended deployment. The low cost requirement of the sensor node precludes using an expensive tamper resistant hardware for sensor physical protection. Thus, the adversary can reprogram the compromised sensors and deviates sensor network functionality. In this paper, we propose two simple software-based remote code attestation schemes for different WSN criterion. Our schemes use different independent memory noise filling techniques called pre-deployment and post-deployment noise filling, and also different communication protocols for attestation purpose. The protocols are well-suited for wireless sensor networks, where external factors , such as channel collision, result in network delay. Hence, the success of our schemes of attestation does not depend on the accurate measurement of the execution time, which is the main drawback of previously proposed wireless sensor network attestation schemes.