In this tutorial, we explore deep learning [32, 53] and neural networks [45] as part of advanced and established machine learning techniques [17–31, 36–44, 47–52, 57], then delve into the evolving intersection of deep learning and social network analysis [33–35]. Such analysis is often used for user emotion recognition [16, 22, 46, 56], which in the context of social media may result in performing sentiment analysis or opinion mining [34]. We highlight how state-of-the-art algorithms—such as Long Short-Term Memory networks (LSTMs) [1], Transformers [2], BERT [3], Graph Neural Networks (GNNs) [4, 9, 15], and Convolutional Neural Networks (CNNs) [5]—are transforming our understanding of complex online social structures. As digital platforms burgeon with vast amounts of textual and relational data, developing robust analytical tools to parse and understand this information is increasingly vital. This session will delve into the application of these advanced deep learning models to dissect both the structural frameworks and the rich text content of online social networks, shedding light on user behavior [6, 14], community dynamics [7, 12], and emerging trends [8, 10]. Additionally, recent advancements such as evidential temporal-aware graph-based event detection [11], deep reinforcement learning approaches for influencer detection [14], and host-based threat detection through provenance graph learning [13] have further expanded analytical capabilities. Through practical demonstrations and case studies, participants will gain firsthand experience in applying these technologies to real-world challenges, thereby enhancing their analytical capabilities and opening up new avenues for research and discovery in digital social interactions and AI-empowered software engineering [54, 55, 58–62].