Mafin: Enhancing Black-Box Embeddings with Model Augmented Fine-Tuning

2024/02/19 by Mingtian Zhang, Ming‐Tian Zhang, Shawn Lan +6 · 1 voice · 1 citation
Computer Science · Engineering · #Topic Modeling #Ferroelectric and Negative Capacitance Devices #Parallel Computing and Optimization Techniques

paper · pdf · doi:10.48550/arxiv.2402.12177

Abstract

Retrieval Augmented Generation (RAG) has emerged as an effective solution for mitigating hallucinations in Large Language Models (LLMs). The retrieval stage in RAG typically involves a pre-trained embedding model, which converts queries and passages into vectors to capture their semantics. However, a standard pre-trained embedding model may exhibit sub-optimal performance when applied to specific domain knowledge, necessitating fine-tuning. This paper addresses scenarios where the embeddings are only available from a black-box model. We introduce Model augmented fine-tuning (Mafin) -- a novel approach for fine-tuning a black-box embedding model by augmenting it with a trainable embedding model. Our results demonstrate that Mafin significantly enhances the performance of the black-box embeddings by only requiring the training of a small augmented model. We validate the effectiveness of our method on both labeled and unlabeled datasets, illustrating its broad applicability and efficiency.

Cited by

• Domain-Specific Data Generation Framework for RAG Adaptation

Discussions

• Fine-Tune Black Box Embedding Models [hn, 10 points, 1 comments]

Related

• Finetune-RAG: Fine-Tuning Language Models to Resist Hallucination in Retrieval-Augmented Generation
• RQ-RAG: Learning to Refine Queries for Retrieval Augmented Generation
• Systematic Knowledge Injection into Large Language Models via Diverse Augmentation for Domain-Specific RAG
• REFINE on Scarce Data: Retrieval Enhancement through Fine-Tuning via Model Fusion of Embedding Models
• LMAR: Language Model Augmented Retriever for Domain-specific Knowledge Indexing