For more details please refer to our github repo: https://github.com/FlagOpen/FlagEmbedding

BGE-M3 (paper, code)

In this project, we introduce BGE-M3, which is distinguished for its versatility in Multi-Functionality, Multi-Linguality, and Multi-Granularity.

• Multi-Functionality: It can simultaneously perform the three common retrieval functionalities of embedding model: dense retrieval, multi-vector retrieval, and sparse retrieval.
• Multi-Linguality: It can support more than 100 working languages.
• Multi-Granularity: It is able to process inputs of different granularities, spanning from short sentences to long documents of up to 8192 tokens.

Some suggestions for retrieval pipeline in RAG

We recommend to use the following pipeline: hybrid retrieval + re-ranking.

• Hybrid retrieval leverages the strengths of various methods, offering higher accuracy and stronger generalization capabilities. A classic example: using both embedding retrieval and the BM25 algorithm. Now, you can try to use BGE-M3, which supports both embedding and sparse retrieval. This allows you to obtain token weights (similar to the BM25) without any additional cost when generate dense embeddings. To use hybrid retrieval, you can refer to Vespa and Milvus.

• As cross-encoder models, re-ranker demonstrates higher accuracy than bi-encoder embedding model. Utilizing the re-ranking model (e.g., bge-reranker, bge-reranker-v2) after retrieval can further filter the selected text.

FAQ

1. Introduction for different retrieval methods

• Dense retrieval: map the text into a single embedding, e.g., DPR, BGE-v1.5
• Sparse retrieval (lexical matching): a vector of size equal to the vocabulary, with the majority of positions set to zero, calculating a weight only for tokens present in the text. e.g., BM25, unicoil, and splade
• Multi-vector retrieval: use multiple vectors to represent a text, e.g., ColBERT.

2. How to use BGE-M3 in other projects?

For embedding retrieval, you can employ the BGE-M3 model using the same approach as BGE. The only difference is that the BGE-M3 model no longer requires adding instructions to the queries.

For hybrid retrieval, you can use Vespa and Milvus.

Evaluation

We provide the evaluation script for MKQA and MLDR

Benchmarks from the open-source community

The BGE-M3 model emerged as the top performer on this benchmark (OAI is short for OpenAI). For more details, please refer to the article and Github Repo

Our results

• Multilingual (Miracl dataset)

• Cross-lingual (MKQA dataset)

• Long Document Retrieval

  • MLDR:
    Please note that MLDR is a document retrieval dataset we constructed via LLM, covering 13 languages, including test set, validation set, and training set. We utilized the training set from MLDR to enhance the model's long document retrieval capabilities. Therefore, comparing baselines with Dense w.o.long(fine-tuning without long document dataset) is more equitable. Additionally, this long document retrieval dataset will be open-sourced to address the current lack of open-source multilingual long text retrieval datasets. We believe that this data will be helpful for the open-source community in training document retrieval models.

  • NarritiveQA:
    

• Comparison with BM25

We utilized Pyserini to implement BM25, and the test results can be reproduced by this script. We tested BM25 using two different tokenizers: one using Lucene Analyzer and the other using the same tokenizer as M3 (i.e., the tokenizer of xlm-roberta). The results indicate that BM25 remains a competitive baseline, especially in long document retrieval.

Acknowledgement

Thanks to the authors of open-sourced datasets, including Miracl, MKQA, NarritiveQA, etc. Thanks to the open-sourced libraries like Tevatron, Pyserini.

Citation

If you find this repository useful, please consider giving a star :star: and citation

@misc{bge-m3,
      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation}, 
      author={Jianlv Chen and Shitao Xiao and Peitian Zhang and Kun Luo and Defu Lian and Zheng Liu},
      year={2024},
      eprint={2402.03216},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}