No Parameter Left Behind: How Distillation and Model Size Affect Zero-Shot Retrieval
This work addresses the challenge of deploying effective retrieval models in real-world applications by revealing that current practices favoring small distilled models may be suboptimal for zero-shot generalization, offering insights for researchers and practitioners in information retrieval.
The paper tackles the problem of zero-shot retrieval generalization by showing that larger model size and early query-document interaction significantly improve performance on unseen domains, with their largest reranker achieving state-of-the-art results in 12 out of 18 BEIR datasets and surpassing the previous best by 3 average points.
Recent work has shown that small distilled language models are strong competitors to models that are orders of magnitude larger and slower in a wide range of information retrieval tasks. This has made distilled and dense models, due to latency constraints, the go-to choice for deployment in real-world retrieval applications. In this work, we question this practice by showing that the number of parameters and early query-document interaction play a significant role in the generalization ability of retrieval models. Our experiments show that increasing model size results in marginal gains on in-domain test sets, but much larger gains in new domains never seen during fine-tuning. Furthermore, we show that rerankers largely outperform dense ones of similar size in several tasks. Our largest reranker reaches the state of the art in 12 of the 18 datasets of the Benchmark-IR (BEIR) and surpasses the previous state of the art by 3 average points. Finally, we confirm that in-domain effectiveness is not a good indicator of zero-shot effectiveness. Code is available at https://github.com/guilhermemr04/scaling-zero-shot-retrieval.git