Deep Learning in the Ubiquitous Human-Computer Interactive 6G Era: Applications, Principles and Prospects

Area of Science:

• Telecommunications Engineering
• Computer Science
• Artificial Intelligence

Background:

• The advent of virtual reality (VR) and augmented reality (AR) heralds a new era of ubiquitous human-centric intelligence.
• Sixth-generation (6G) wireless systems are crucial for seamless human-computer interaction in this future.
• 6G promises significant performance enhancements over previous generations, enabling advanced applications.

Purpose of the Study:

• To address the challenges hindering the development of 6G systems.
• To explore the application of deep learning techniques to overcome these 6G challenges.
• To provide a systematic review of deep learning solutions for 6G.

Main Methods:

• Reviewing representative deep learning solutions for six key 6G challenges: Terahertz/millimeter-wave communication, low latency/high reliability, energy efficiency, security, edge computing, and service heterogeneity.
• Analyzing the principles of applying deep learning to specific 6G issues.
• Investigating the role of deep reinforcement learning and addressing data scarcity in 6G.

Main Results:

• Deep learning provides effective alternatives to traditional analytical methods for complex 6G problems.
• Deep reinforcement learning is identified as vital for 6G systems.
• Solutions for training data scarcity and insights into the synergy between traditional methods and deep learning in 6G are presented.
• Frequently used deep learning techniques for 6G are identified.

Conclusions:

• Deep learning is essential for realizing the full potential of 6G human-centric applications.
• Addressing challenges in communication, efficiency, security, and computing through AI is critical for 6G.
• Future research should focus on open problems and advanced deep learning applications in 6G networks.