Due to the increasing severity of information security threats that affect national security and society, the "Future Network Security" comprehensive project is crucial. The main goal of this project is to strengthen AI network security and post-quantum cryptography, as well as improve national information security and contribute to academic research. The project is expected to include six subprojects: "Detecting Hidden Network Attacks Through Data Correlation and Source Analysis" uses the MITRE ATT&CK framework to develop deep learning algorithms for an effective intrusion detection system. "Enhancing Malicious Behavior Discovery Using Network Threat Intelligence and Generated Malware" focuses on proactive malware defense by integrating the MITRE ATT&CK framework and CTI reports to generate TTP malware. "Practice, Theory, and Defense of Adversarial Network Attacks" investigates the robustness of AI models against adversarial attacks. "Robust Multimodal Deepfake Detection and Mitigation" detects multimodal deepfake images and videos by designing robust deepfake detection algorithms. "Zero Trust Federated Learning" proposes a zero trust federated learning framework, researching the security, privacy, robustness, verifiability, and fairness of federated learning. "Post-Quantum Cryptography" tracks the NIST post-quantum cryptography standardization process, assisting in developing more secure cryptographic systems to defend against quantum attacks and post-quantum migration. This project has enormous potential impact and benefits, including improving national information security and contributing to academic research. By focusing on confidentiality, integrity, and availability, the project will provide comprehensive protection. It will promote the long-term development of the network security industry and help build an intelligent and secure nation.