This presentation provides an in-depth look at RAPIEnet, an innovative Korean real-time Ethernet technology. We will begin with an overview of RAPIEnet and its origins, followed by a comprehensive exploration of its applications across diverse sectors, demonstrated with engaging case studies. We will delve into RAPIEnet's unique capabilities and the tangible benefits they bring, focusing on how this breakthrough technology can improve competitiveness within both domestic and international landscapes. A special focus will be given to companies like LS ELECTRIC, as we discuss how RAPIEnet can help such enterprises secure a strong foothold in the global market. Ultimately, we aim to deliver a profound understanding of RAPIEnet's potential, shedding light on the exciting future possibilities this technology holds.

DetNet [RFC8655] provides the capability to carry specified unicast or multicast data flows with extremely low data loss rates and bounded end-to-end latency within a network domain. DetNet may be used in industrial scenarios, 5G transport network and media transport; this presentation will introduce DetNet's development history, key technology solutions and the current progress of the standard in IETF.

3GPP Release 18 has been developing UPF enhancement for Exposure and SBA (UPEAS) to more effectively integrate UPF into the 5GC SBA. This presentation introduces UPEAS related key technologies and benefits as well as the key features of the 5G/6G user plane from the CMCC side.

As we move beyond using the knowledge and creative capabilities of Large Language Models (LLMs) like ChatGPT, research and development are underway to establish an AI ecosystem centered around LLMs. This presentation will introduce a system that leverages the remarkable performance and flexibility of LLMs for general tasks, integrating a variety of tools through collaboration with task-specific AIs to address user requests. Our approach, by integrating the strengths of specific AI with the general capabilities of LLMs, creates a more holistic, robust, and adaptable solution to meet diverse user requirements. Further, we will delve into the concept of autonomous agents that can contribute to the improvement and optimization of the system's decision-making process. These agents, working in sync with LLMs, represent the next step in our journey towards a truly intelligent and responsive AI ecosystem. Join us as we explore the future of AI, its applications, and how LLM-based agents are reshaping the landscape of AI technology.

With a growing interest in sleep monitoring at home, sound-based sleep staging with deep learning has emerged as a potential solution. However, collecting labeled data is restrictive in the home environments due to the inconvenience of installing medical equipment at home. To handle this, we propose novel training approaches using accessible real-world sleep sound data. Our key contributions include a new semi-supervised learning technique that considers the time-series nature of sleep sound. Our model was evaluated on various datasets including a labeled home sleep sound dataset and the public PSG-Audio dataset, demonstrating the robustness and generalizability of our model across real-world scenarios.

Toward the 6G mobile network, KDDI has proposed a 6G RAN architecture, named user-centric RAN, to provide personalized radio signals for each user. Based on the promising idea of convergence between RAN and CORE networks, we are proposing the 6G mobile core system called user-centric core to cooperate with user-centric RAN, which is a highly efficient, automated, and customizable mobile network which satisfies the diverse requirements of each user/service. From the U-Plane perspective, we believe that the conventional UPF processing, which binds the forwarding process of one IP flow to one CPU core, may become a bottleneck in achieving high-capacity communication requirements in 6G, and we are studying the feasibility of a UPF acceleration method by assigning multiple CPU cores to one IP flow. Our presentation introduces the key technologies to realize them, i.e., "Full-SBA," "Lightweight," "Component," and "multi-CPU core processing UPF."