No Parameter Left Behind: How Distillation and Model Size Affect Zero-Shot Retrieval

TL;DR

This paper investigates how model size and early query-document interaction influence zero-shot retrieval performance, showing larger models and rerankers significantly improve generalization across diverse datasets.

Contribution

It demonstrates that increasing model size and using rerankers enhances zero-shot retrieval, surpassing previous state-of-the-art results on the BEIR benchmark.

Findings

Larger models yield greater gains in unseen domains.

Rerankers outperform dense models of similar size.

Largest reranker achieves state-of-the-art on 12 of 18 datasets.

Abstract

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