Deep learning models have been used in creating various effective image classification applications. However, they are vulnerable to adversarial attacks that seek to misguide the models into predicting incorrect classes. Our study of major adversarial attack models shows that they all specifically target and exploit the neural networking structures in their designs. This understanding led us to develop a hypothesis that most classical machine learning models, such as random forest (RF), are immune to adversarial attack models because they do not rely on neural network design at all. Our experimental study of classical machine learning models against popular adversarial attacks supports this hypothesis. Based on this hypothesis, we propose a new adversarial-aware deep learning system by using a classical machine learning model as the secondary verification system to complement the primary deep learning model in image classification. Although the secondary classical machine learning model has less accurate output, it is only used for verification purposes, which does not impact the output accuracy of the primary deep learning model, and, at the same time, can effectively detect an adversarial attack when a clear mismatch occurs. Our experiments based on the CIFAR-100 dataset show that our proposed approach outperforms current state-of-the-art adversarial defense systems.

Preserving a secure and morally safe online environment on social media is a challenging task. It is essential to find immoral or unsuitable information in user-generated postings to safeguard users and enforce community standards. Various Natural Language Processing (NLP) approaches are being employed to detect subtle immoral posts; however, there remains a research gap due to the semantic and contextual complexity of natural language. To bridge this gap, this work proposes the use of a Knowledge Graph (KG) for entity recognition and the extraction of semantic relationships in Social Network (SN) posts. By doing so, the KG helps provide a deeper contextual understanding, enabling the detection of negative interactions between entities that are often present in immoral content. KG allows us to extract these associations from the text, enabling the model to recognize language that leads to immoral behavior. By utilizing a KG, the model can more easily identify connections between entities, verify statements made in postings, and classify material more precisely. GloVe (Global Vector) word embedding is used to transform the enriched text data into numerical representations. An attention-based Bidirectional Long Short-Term Memory (BiLSTM) network performs the classification task. The BiLSTM concurrently analyses the input sequence in both directions, enabling the network to recognize not only the context that is present at the moment but also the context in which each word in the sequence comes before and after. To validate the model performance, we used benchmark datasets Self-Annotated Reddit Corpus (SARC), and Hate Evaluation (HatEval) dataset. We achieved a higher F1-score of 82.79% and 84.06% on both datasets and outperformed state-of-the-art works.

Deep learning models have been used in creating various effective image classification applications. However, they are vulnerable to adversarial attacks that seek to misguide the models into predicting incorrect classes. Our study of major adversarial attack models shows that they all specifically target and exploit the neural networking structures in their designs. This understanding makes us develop a hypothesis that most classical machine learning models, such as Random Forest (RF), are immune to adversarial attack models because they do not rely on neural network design at all. Our experimental study of classical machine learning models against popular adversarial attacks supports this hypothesis. Based on this hypothesis, we propose a new adversarial-aware deep learning system by using a classical machine learning model as the secondary verification system to complement the primary deep learning model in image classification. Although the secondary classical machine learning model has less accurate output, it is only used for verification purposes, which does not impact the output accuracy of the primary deep learning model, and at the same time, can effectively detect an adversarial attack when a clear mismatch occurs. Our experiments based on CIFAR-100 dataset show that our proposed approach outperforms current state-of-the-art adversarial defense systems.