KaLM-Embedding: Superior Training Data Brings A Stronger Embedding Model
This work addresses the need for better multilingual embedding models in retrieval-augmented generation, though it is incremental as it builds on existing techniques with a focus on data quality.
The paper tackles the problem of improving embedding models for retrieval-augmented generation by focusing on training data quality, resulting in KaLM-Embedding, which outperforms comparable models on the MTEB benchmark across multiple languages.
As retrieval-augmented generation prevails in large language models, embedding models are becoming increasingly crucial. Despite the growing number of general embedding models, prior work often overlooks the critical role of training data quality. In this work, we introduce KaLM-Embedding, a general multilingual embedding model that leverages a large quantity of cleaner, more diverse, and domain-specific training data. Our model has been trained with key techniques proven to enhance performance: (1) persona-based synthetic data to create diversified examples distilled from LLMs, (2) ranking consistency filtering to remove less informative samples, and (3) semi-homogeneous task batch sampling to improve training efficacy. Departing from traditional BERT-like architectures, we adopt Qwen2-0.5B as the pre-trained model, facilitating the adaptation of auto-regressive language models for general embedding tasks. Extensive evaluations of the MTEB benchmark across multiple languages show that our model outperforms others of comparable size, setting a new standard for multilingual embedding models with <1B parameters.