MixLoRA-DSI: Dynamically Expandable Mixture-of-LoRA Experts for Rehearsal-Free Generative Retrieval over Dynamic Corpora
This addresses the challenge of continual model updates in retrieval systems for applications with evolving data, though it is incremental as it builds on existing LoRA and mixture-of-experts methods.
The paper tackles the problem of efficiently updating generative retrieval models for dynamic corpora without full retraining, achieving performance gains with minimal parameter overhead and lower training costs on datasets like NQ320k and MS MARCO Passage.
Continually updating model-based indexes in generative retrieval with new documents remains challenging, as full retraining is computationally expensive and impractical under resource constraints. We propose MixLoRA-DSI, a novel framework that combines an expandable mixture of Low-Rank Adaptation experts with a layer-wise out-of-distribution (OOD)-driven expansion strategy. Instead of allocating new experts for each new corpus, our proposed expansion strategy enables sublinear parameter growth by selectively introducing new experts only when significant number of OOD documents are detected. Experiments on NQ320k and MS MARCO Passage demonstrate that MixLoRA-DSI outperforms full-model update baselines, with minimal parameter overhead and substantially lower training costs.