Biju Issac

Smart grid demand response (DR) systems face escalating threats from False Data Injection Attacks (FDIAs), which corrupt load forecasts to distort real-time pricing, destabilize supply– demand equilibrium, and precipitate cascading grid failures. These intrusions manifest as multi-scale temporal anomalies from transient spikes to prolonged distortions in severely imbalanced datasets. We propose an adaptive deep learning architecture integrating: (i) a variance-adaptive 1D convolutional neural network (CNN) with dynamic kernel scaling to magnify anomalies while attenuating noise; (ii) a temporal-attention bidirectional LSTM (BiLSTM) to resolve forward–backward dependencies; and (iii) Bayesian optimization (BO) with Matérn kernels to jointly optimize accuracy–latency trade-offs under strict computational constraints. An Enhanced SMOTE-RUS scheme synthesizes minority-class attacks while preserving appliance-level periodicity. Evaluated on 377k half-hourly forecasts from 290 U.S. households, the framework attains 99.26% accuracy, 93.16% F1-score, and 1.2 ms per-sample inference, exceeding the strongest baseline by 21.4%.

Since the Internet of Things (IoT) is widely adopted using Android applications, detecting malicious Android apps is essential. In recent years, Android graph-based deep learning research has proposed many approaches to extract relationships from applications as graphs to generate graph embeddings. First, we demonstrate the effectiveness of graph-based classification using a Graph Neural Network (GNN)-based classifier to generate API graph embeddings. The graph embeddings are combined with Permission and Intent features to train multiple machine learning and deep learning models for Android malware detection. The proposed classification approach achieves an accuracy of 98.33 percent on the CICMaldroid dataset and 98.68 percent on the Drebin dataset. However, graph-based deep learning models are vulnerable, as attackers can add fake relationships to evade detection by the classifier. Second, we propose a Generative Adversarial Network (GAN)-based attack algorithm named VGAE-MalGAN targeting graph-based GNN Android malware classifiers. The VGAE-MalGAN generator produces adversarial malware API graphs, while the VGAE-MalGAN substitute detector attempts to mimic the target detector. Experimental results show that VGAE-MalGAN can significantly reduce the detection rate of GNN-based malware classifiers. Although the model initially fails to detect adversarial malware, retraining with generated adversarial samples improves robustness and helps mitigate adversarial attacks.

As AI-generated images become increasingly photorealistic, distinguishing them from natural images poses a growing challenge. This paper presents a robust detection framework that leverages multiple uncertainty measures to decide whether to trust or reject a model's predictions. We focus on three complementary techniques: Fisher Information, which captures the sensitivity of model parameters to input variations; entropy-based uncertainty from Monte Carlo Dropout, which reflects predictive variability; and predictive variance from a Deep Kernel Learning framework using a Gaussian Process classifier. To integrate these diverse uncertainty signals, Particle Swarm Optimisation is used to learn optimal weightings and determine an adaptive rejection threshold. The model is trained on Stable Diffusion-generated images and evaluated on GLIDE, VQDM, Midjourney, BigGAN, and StyleGAN3, each introducing significant distribution shifts. While standard metrics such as prediction probability and Fisher-based measures perform well in distribution, their effectiveness degrades under shift. In contrast, the Combined Uncertainty measure consistently achieves an incorrect rejection rate of approximately 70 percent on unseen generators, successfully filtering most misclassified AI samples. Although the system occasionally rejects correct predictions from newer generators, this conservative behaviour is acceptable, as rejected samples can support retraining. The framework maintains high acceptance of accurate predictions for natural images and in-domain AI data. Under adversarial attacks using FGSM and PGD, the Combined Uncertainty method rejects around 61 percent of successful attacks, while GP-based uncertainty alone achieves up to 80 percent. Overall, the results demonstrate that multi-source uncertainty fusion provides a resilient and adaptive solution for AI-generated image detection.

As AI-generated images become increasingly photorealistic, distinguishing them from natural images poses a growing challenge. This paper presents a robust detection framework that leverages multiple uncertainty measures to decide whether to trust or reject a model’s predictions. We focus on three complementary techniques: Fisher Information, which captures the sensitivity of model parameters to input variations; entropy-based uncertainty from Monte Carlo (MC) Dropout, which reflects predictive variability; and predictive variance from a Deep Kernel Learning (DKL) framework using a Gaussian Process (GP) classifier. To integrate these diverse uncertainty signals, we employ Particle Swarm Optimisation (PSO) to learn optimal weightings and determine an adaptive rejection threshold. The model is trained on Stable Diffusion-generated images and evaluated on GLIDE, VQDM, Midjourney, BigGAN and StyleGAN3 each presenting significant distribution shifts. While standard metrics like prediction probability and Fisher-based measures perform well in distribution, they degrade under shift. In contrast, the Combined Uncertainty measure consistently achieves an incorrect rejection rate of approximately 70% on unseen generators, successfully filtering out most misclassified AI samples. Although the system occasionally rejects correct predictions from newer generators, this conservative behaviour remains acceptable, as rejected synthetic samples serve as valuable input for retraining. Crucially, it maintains high acceptance of accurate predictions for natural images and in-domain AI data. Under adversarial attacks (FGSM and PGD), the Combined Uncertainty method rejects around 61% of successful attacks, while the GP-based uncertainty alone achieves up to 80%. Notably, the Combined approach maintains strong selectivity, rarely rejecting correct predictions. Overall, our findings highlight the benefit of multi-source uncertainty fusion for resilient and adaptive AI-generated image detection.

Ransomware poses a serious and evolving threat, demanding detection methods that can adapt to new attack vectors while maintaining transparency and reliability. This study proposes a comprehensive framework that integrates data augmentation, explainable artificial intelligence, and uncertainty quantification to address key challenges in ransomware detection. By leveraging synthetic data generation techniques, the approach mitigates class imbalance and captures varied ransomware behaviours. Simultaneously, explainable AI methods shed light on model decisions, enhancing interpretability and building trust among cybersecurity professionals. An uncertainty-aware component flags ambiguous predictions, allowing for targeted manual reviews and minimising incorrect classifications. Experiments on multiple ransomware datasets show the framework’s ability to maintain high detection rates, even under adversarial conditions. By combining RanSAP and RDset datasets, the framework achieves marked performance improvements. When SMOTE was applied, Random Forest reached an F1-score of 0.9963, while a CNN with Monte Carlo Dropout attained 0.9906. Further incorporating CT-GAN boosted the CNN’s F1-score to 0.9978, underscoring the robustness of our approach. The results suggest that combining robust data augmentation, interpretability, and uncertainty handling offers a practical avenue for deploying reliable ransomware detection systems in real-world environments.

Current mobile payment schemes gather detailed information about purchases customers make. This data can then be used to infer a customer’s spending behaviour, potentially violating their privacy. To tackle this problem, we propose an untraceable mobile payment scheme that strikes a better balance, preserving user privacy while allowing the Third-Party Service Provider (TPSP) to collect necessary information such as card details and transaction amount for regulatory compliance. Our scheme offers untraceability for legitimate users from malicious adversaries and curious TPSPs using cryptographic primitives such as partially blind signatures, zero-knowledge proofs, and identity-based signatures. It also guarantees that only authorised TPSPs can issue valid payment tokens, and even with limited data, the TPSP can still prevent dishonest customers/merchants from double-spending a payment token. We also propose a comprehensive evaluation framework to assess the untraceable payment schemes against seven key criteria such as untraceability, exculpability—merchant double-spending, exculpability—customer double-spending, unforgeability, confidentiality, message authenticity, efficiency, and regulatory compliance. We rigorously benchmark the security and privacy of our proposed payment scheme against this framework and other established schemes. Furthermore, we formally verify these properties using complexity-based analysis and Proverif modelling.

The project proposes an AI-driven framework to enhance malware and botnet detection in smart buildings, employing blockchain and federated learning. It aims to overcome the limitations of conventional security methods by introducing a more dynamic and intelligent detection system. The initiative seeks to improve IoT security, establishing new standards for protecting smart building infrastructures against evolving cyber threats. Through the integration of blockchain and federated learning, the project addresses data security and privacy challenges, ensuring robust detection capabilities across various IoT environments.

Botnets are computer networks controlled by malicious actors that present significant cybersecurity challenges. They autonomously infect, propagate, and coordinate to conduct cybercrimes, necessitating robust detection methods. This research addresses the sophisticated adversarial manipulations posed by attackers, aiming to undermine machine learning-based botnet detection systems. We introduce a flow-based detection approach, leveraging machine learning and deep learning algorithms trained on the ISCX and ISOT datasets. The detection algorithms are optimized using the Genetic Algorithm and Particle Swarm Optimization to obtain a baseline detection method. The Carlini & Wagner (C&W) attack and Generative Adversarial Network (GAN) generate deceptive data with subtle perturbations, targeting each feature used for classification while preserving their semantic and syntactic relationships, which ensures that the adversarial samples retain meaningfulness and realism. An in-depth analysis of the required L2 distance from the original sample for the malware sample to misclassify is performed across various iteration checkpoints, showing different levels of misclassification at different L2 distances of the Pertrub sample from the original sample. Our work delves into the vulnerability of various models, examining the transferability of adversarial examples from a Neural Network surrogate model to Tree-based algorithms. Subsequently, models that initially misclassified the perturbed samples are retrained, enhancing their resilience and detection capabilities. In the final phase, a conformal prediction layer is integrated, significantly rejecting incorrect predictions, of 58.20 % in the ISCX dataset and 98.94 % in the ISOT dataset.