GENAI: RAG USE CASES WITH VECTOR DB TO SOLVE THE LIMITATIONS OF LLMS
Keywords:
Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Vector Databases (Vector DB), Semantic EmbeddingsAbstract
This research delves into combining Retrieval Augmented Generation (RAG) with Vector Databases (Vector DB) to tackle the challenges faced by Language Models (LLMs) in Generative AI scenarios. Despite the progress made by LLMs in understanding and generating language issues such as data, hallucination and incorporating domain specific details persist. Our innovative method utilizes the semantically robust features of Vector DBs in conjunction with the RAG framework to improve aspects of LLM performance. By integrating time relevant information retrieved from Vector DBs LLMs can produce more precise, current, and targeted content. We outline the procedures involved in gathering data from sources creating embeddings and assigning metadata to establish a repository that significantly enhances LLM generative capabilities. Our results suggest that this approach not only overcomes LLM limitations but also opens up opportunities for their utilization, in fields requiring accuracy and timeliness.
References
An, Fengwei et al. “VLSI realization of learning vector quantization with hardware/software co-design for different applications.” Japanese Journal of Applied Physics 54 (2015): n. pag.
Gao, Yunfan et al. “Retrieval-Augmented Generation for Large Language Models: A Survey.” ArXiv abs/2312.10997 (2023): n. pag.
Li, Ziqi and A. Stewart Fotheringham. “Computational improvements to multi-scale geographically weighted regression.” International Journal of Geographical Information Science 34 (2019): 1378 - 1397.
Zhang, Xiaowei et al. “Sparse Uncorrelated Linear Discriminant Analysis for Undersampled Problems.” IEEE Transactions on Neural Networks and Learning Systems 27 (2016): 1469-1485.
Moll, Simon et al. “Multi-dimensional Vectorization in LLVM.” WPMVP'19 (2019).
Bragin, Mikhail A. et al. “A Scalable Solution Methodology for Mixed-Integer Linear Programming Problems Arising in Automation.” IEEE Transactions on Automation Science and Engineering 16 (2019): 531-541.
Zhu, Junan and Dror Baron. “Performance Limits with Additive Error Metrics in Noisy Multimeasurement Vector Problems.” IEEE Transactions on Signal Processing 66 (2018): 5338-5348.
Silva, Danilo Avilar and Ajalmar R. da Rocha Neto. “A Genetic Algorithms-Based LSSVM Classifier for Fixed-Size Set of Support Vectors.” International Work-Conference on Artificial and Natural Neural Networks (2015).
