Ahmed Dawod Mohammed Ibrahum

Mr. Ahmed Dawod Mohammed Ibrahum


Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 November 2024 in Artificial Intelligence Review
Verified
Journal Article
Open access
PDF file
AIAhmed Dawod Mohammed Ibrahum
Ahmed Dawod Mohammed Ibrahum
MHManzoor Hussain
Manzoor Hussain
JHJang-Eui Hong
Jang-Eui Hong
The integration of Deep Learning (DL) algorithms in Autonomous Vehicles (AVs) has revolutionized their precision in navigating various driving scenarios, ranging from anti-fatigue safe driving to intelligent route planning. Despite their proven effectiveness, concerns regarding the safety and reliability of DL algorithms in AVs have emerged, particularly in light of the escalating threat of adversarial attacks, as emphasized by recent research. These digital or physical attacks present formidable challenges to AV safety, relying extensively on collecting and interpreting environmental data through integrated sensors and DL. This paper addresses this pressing issue through a systematic survey that meticulously explores robust adversarial attacks and defenses, specifically focusing on DL in AVs from a safety perspective. Going beyond a review of existing research papers on adversarial attacks and defenses, the paper introduces a safety scenarios taxonomy matrix Inspired by SOTIF designed to augment the safety of DL in AVs. This matrix categorizes safety scenarios into four distinct areas and classifies attacks into those areas in three scenarios, along with two defense scenarios. Furthermore, the paper investigates the testing and evaluation measurements critical for assessing attacks in the context of DL for AVs. It further explores the dynamic landscape of datasets and simulation platforms. This contribution significantly enriches the ongoing discourse surrounding the assurance of safety and reliability in autonomous vehicles, especially in the face of continually evolving adversarial challenges.

6 November 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
AIAhmed Dawod Mohammed Ibrahum
Ahmed Dawod Mohammed Ibrahum
ZSZhengyu Shang
Zhengyu Shang
JHJang-Eui Hong
Jang-Eui Hong
Kolmogorov–Arnold Networks (KANs) are a novel class of neural network architectures based on the Kolmogorov–Arnold representation theorem, which has demonstrated potential advantages in accuracy and interpretability over Multilayer Perceptron (MLP) models. This paper comprehensively evaluates the robustness of various KAN architectures—including KAN, KAN-Mixer, KANConv_KAN, and KANConv_MLP—against adversarial attacks, which constitute a critical aspect that has been underexplored in current research. We compare these models with MLP-based architectures such as MLP, MLP-Mixer, and ConvNet_MLP across three traffic sign classification datasets: GTSRB, BTSD, and CTSD. The models were subjected to various adversarial attacks (FGSM, PGD, CW, and BIM) with varying perturbation levels and were trained under different strategies, including standard training, adversarial training, and Randomized Smoothing. Our experimental results demonstrate that KAN-based models, particularly the KAN-Mixer, exhibit superior robustness to adversarial attacks compared to their MLP counterparts. Specifically, the KAN-Mixer consistently achieved lower Success Attack Rates (SARs) and Degrees of Change (DoCs) across most attack types and datasets while maintaining high accuracy on clean data. For instance, under FGSM attacks with ϵ=0.01, the KAN-Mixer outperformed the MLP-Mixer by maintaining higher accuracy and lower SARs. Adversarial training and Randomized Smoothing further enhanced the robustness of KAN-based models, with t-SNE visualizations revealing more stable latent space representations under adversarial perturbations. These findings underscore the potential of KAN architectures to improve neural network security and reliability in adversarial settings.

28 August 2025 in IEEE Access
Verified
Journal Article
Open access
MHManzoor Hussain
Manzoor Hussain
ZSZhengyu Shang
Zhengyu Shang
AIAhmed Dawod Mohammed Ibrahum
Ahmed Dawod Mohammed Ibrahum
JHJang-Eui Hong
Jang-Eui Hong
Adversarial training is a widely used method to improve the robustness of deep learning models in various applications. Although adversarial training enhances the robustness of the target model, it also suffers from an accuracy versus robustness trade-off, meaning that while improving robustness, the accuracy of the target model decreases. Moreover, they also suffer from catastrophic overfitting, where these methods offer reasonable robustness against specific adversarial attacks, but they become ineffective against others. Thus, to solve these issues, we proposed Prediction Error-Guided Adversarial Training (PEGAT). In this method, we first evaluate the normally trained target models against adversarial samples and then calculate class-wise prediction error in terms of loss in the F1-score, which we refer to as PEGAT scores. In the second phase, use the PEGAT scores as guidance to assign weights during the adversarial training process. In this phase, PEGAT adaptively assigns higher weights to classes with high PEGAT scores and lower weights to classes with low PEGAT scores. The PEGAT guides the models to focus on those classes where they struggle to maintain a higher F1 score in the first phase. By doing so, not only is the adversarial robustness of the target model increased, but also the issues such as the accuracy versus robustness tradeoff, catastrophic overfitting, and generalizability are addressed. Evaluating the PEGAT on classification models used in intelligent autonomous vehicle perception systems, we found that the technique could improve the robustness of WideResNet18, WideResNet34, and WideResNet50 by 16%, 6%, and 5%, respectively, without suffering from the aforementioned issues.

9 February 2023
Verified
Conference Paper
Open access
AIAhmed Dawod Mohammed Ibrahum
Ahmed Dawod Mohammed Ibrahum
munkhdalai lhagiimunkhdalai lhagii
munkhdalai lhagii
KPKwang Ho Park
Kwang Ho Park
Keun Ho RyuKeun Ho Ryu
Keun Ho Ryu
VPVan Huy Pham
Van Huy Pham
Non-Fungible Tokens (NFTs) are digital assets based on a blockchain and those are characterized as unique cryptographic tokens and non-interchangeable. To date, research into the NFT marketplace has been relatively limited. As it is an emerging platform with many unique elements, The NFT market has been impacted due to recent fluctuations in crypto-asset markets more broadly. This current bear market cycle has shed light on concerns around the value of NFTs, profit-based motivation, and environmental sustainability. However, periods of volatility and cyclicality are to be expected with any nascent technology as it develops a product-market fit. consequently, the appraisal of real-price for NFT collections is essential for individual financial security and investment making. In this study, we evaluate the machine learning algorithms to appraise their real-price based on NFT item's characteristics, market event information, and their rarity score data acquired by retrieved from the biggest marketplace OpenSea. Furthermore, the procedures were applied to meet the objectives of this study we built prediction models based on various machine-learning algorithms ranging from Random Forest, XGBoost, SVM, Lasso, ElasticNet, Ridge, Linear Polynomial Regression, TabNet, CatBoost, and LightGBM models. From the results, LightGBM regression model outperformed the other by RMSE around 0.905. The best R2 is only found in this model, which has a value of 0.917.

9 June 2026 in IEEE Access
Verified
Journal Article
Open access
AIAhmed Dawod Mohammed Ibrahum
Ahmed Dawod Mohammed Ibrahum
JHJang-Eui Hong
Jang-Eui Hong
Adversarial training remains the most effective empirical defense against adversarial examples, yet it is hindered by the high cost of inner maximization, the clean–robust accuracy trade-off, and the underutilization of information contained in misclassified or near-boundary adversarial samples.We propose DP2O, a preference-guided framework that converts adversarial evaluations into structured preference pairs and operationalizes Direct Preference Optimization (DPO) for vision robustness. Specifically, we are the first to import the implicit-reward, reference-anchored DPO formulation into the adversarial robustness setting for vision models, andwe do so through a dual-pool construction in which adversarial failures formunstable pairs (U-pairs) preferring the ground-truth label over the model’s incorrect prediction, while adversarial successes form resilient pairs (R-pairs) preferring the ground-truth label over the runner-up class. A hybrid objective then combinesDPOon these typed pairs with cross-entropy on resilient examples.Across CIFAR-10, CIFAR-100, and STL-10 using ResNet-18 and a diverse attack suite (FGSM, PGD8,20, PGD64,40, EOTPGD, Jitter, andAutoAttack), DP2Oconsistently improves adversarial robustness over strong baselines while maintaining competitive clean accuracy. On CIFAR-10, the hybrid model improves PGD8,20 robustness from 0.4853 (TRADES) to 0.5846, increases FGSM robustness to 0.6718, and achieves 0.4934 under AutoAttack while maintaining 0.9125 clean accuracy. Similar gains are observed on CIFAR-100 and STL-10, where DP2O improves robustness across both iterative and adaptive attacks. These results demonstrate that reframing adversarial robustness as preference learning effectively reuses adversarial outcomes as supervision and improves the clean–robust trade-off without requiring additional critic models.

3 April 2026
Verified
Book Chapter
Closed access
AIAhmed Dawod Mohammed Ibrahum
Ahmed Dawod Mohammed Ibrahum
CDCong Ding
Cong Ding
munkhdalai lhagiimunkhdalai lhagii
munkhdalai lhagii
KPKwang Ho Park
Kwang Ho Park
VPVan Huy Pham
Van Huy Pham
Keun Ho RyuKeun Ho Ryu
Keun Ho Ryu