BROWSE
Detail
Conference Paper
Towards a personalized LLM-based daily edge memory aid
Cited 0 time in 
Share 
Share
- Authors
- Sangil Ha, Dongbeom Ko, Joon-Ho Lim, Sungjoo Kang, Hyeon Soo Kim
- Issue Date
- 2024-10
- Citation
- International Conference on Information and Communication Technology Convergence (ICTC) 2024, pp.860-865
- Publisher
- IEEE
- Language
- English
- Type
- Conference Paper
- Abstract
- This paper introduces a personalized memory aid system that combines a high-performance cloud-based large language model (LLM) with a low-power edge device running a small language model (sLM) to enhance user productivity by processing and summarizing daily conversations in real-time. The system optimizes efficiency and user experience while ensuring privacy through localized data handling. By combining cloud and edge resources, the system provides scalable, real-time memory support without compromising data security. The research highlights the system's architecture, its role in protecting user data, and its potential to seamlessly integrate into daily life, offering a sustainable and efficient approach to personalized artificial intelligence applications.
- KSP Keywords
- Data handling, Data security, Edge devices, Efficient approach, High performance, Language Model, Memory support, User data, User experience, artificial intelligence applications, cloud-based