M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation

TL;DR

M3-Embedding is a versatile multilingual and multi-functional text embedding model capable of handling various retrieval tasks and input granularities, trained with a novel self-knowledge distillation method, achieving state-of-the-art results.

Contribution

The paper introduces M3-Embedding, a novel multi-lingual, multi-functionality, and multi-granularity embedding model with a unique self-knowledge distillation training approach.

Findings

Achieves state-of-the-art results on multilingual retrieval benchmarks.

Supports over 100 languages with high retrieval accuracy.

Handles inputs from short sentences to long documents up to 8,192 tokens.

Abstract

In this paper, we introduce a new embedding model called M3-Embedding, which is distinguished for its versatility in \textit{Multi-Linguality}, \textit{Multi-Functionality}, and \textit{Multi-Granularity}. It provides a uniform support for the semantic retrieval of more than 100 working languages. It can simultaneously accomplish the three common retrieval functionalities: dense retrieval, multi-vector retrieval, and sparse retrieval. Besides, it is also capable of processing inputs of different granularities, spanning from short sentences to long documents of up to 8,192 tokens. The effective training of M3-Embedding presents a series of technical contributions. Notably, we propose a novel self-knowledge distillation approach, where the relevance scores from different retrieval functionalities can be integrated as the teacher signal to enhance the training quality. We also optimize the batching strategy, which enables a large batch size and high training throughput to improve the discriminativeness of embeddings. M3-Embedding exhibits a superior performance in our experiment, leading to new state-of-the-art results on multilingual, cross-lingual, and long-document retrieval benchmarks.