Simple and Efficient ways to Improve REALM
This work provides incremental but practical improvements for researchers and practitioners using dense retrieval systems for open-domain QA.
The authors tackled the problem of improving REALM's performance on open-domain QA tasks by optimizing its fine-tuning process, finding that simple improvements in training, supervision, and inference setups significantly boost results. Their best model, REALM++, achieves ~5.5% absolute accuracy improvement over baselines and matches the performance of models with 3x more parameters.
Dense retrieval has been shown to be effective for retrieving relevant documents for Open Domain QA, surpassing popular sparse retrieval methods like BM25. REALM (Guu et al., 2020) is an end-to-end dense retrieval system that relies on MLM based pretraining for improved downstream QA efficiency across multiple datasets. We study the finetuning of REALM on various QA tasks and explore the limits of various hyperparameter and supervision choices. We find that REALM was significantly undertrained when finetuning and simple improvements in the training, supervision, and inference setups can significantly benefit QA results and exceed the performance of other models published post it. Our best model, REALM++, incorporates all the best working findings and achieves significant QA accuracy improvements over baselines (~5.5% absolute accuracy) without any model design changes. Additionally, REALM++ matches the performance of large Open Domain QA models which have 3x more parameters demonstrating the efficiency of the setup.