Learned Sparse IR models, such as SPLADE, offer an excellent efficiency-effectiveness tradeoff. However, they rely on the underlying backbone vocabulary, which might hinder performance (polysemicity and synonymy) and pose a challenge for multi-lingual and multi-modal usages. To solve this limitation, we propose to replace the backbone vocabulary with a latent space of semantic concepts learned using Sparse Auto-Encoders (SAE). Throughout this paper, we study the compatibility of these 2 concepts, explore training approaches, and analyze the differences between our SAE-SPLADE model and traditional SPLADE models. Our experiments demonstrate that SAE-SPLADE achieves retrieval performance comparable to SPLADE on both in-domain and out-of-domain tasks while offering improved efficiency.

翻译：学习型稀疏信息检索模型（如SPLADE）在效率与效果之间实现了优异的平衡。然而，这类模型依赖于底层骨干词汇表，这可能因多义词和同义词现象而制约性能，并为多语言和多模态应用带来挑战。为解决这一局限，我们提出用基于稀疏自编码器学习的语义概念隐空间替代传统骨干词汇表。本文研究了这两种概念的兼容性，探索了训练方法，并分析了SAE-SPLADE模型与传统SPLADE模型的差异。实验表明，SAE-SPLADE在领域内和跨领域检索任务中均能达到与SPLADE相当的检索性能，同时具有更优的效率。