Chain-of-Skills: A Configurable Model for Open-domain Question Answering
This work addresses the need for more adaptable and scalable retrieval models in knowledge-intensive tasks like open-domain question answering, offering a configurable solution that improves performance across multiple datasets.
The paper tackles the problem of limited transferability and scalability in retrieval models for open-domain question answering by proposing a modular retriever with reusable skills, which outperforms recent self-supervised retrievers in zero-shot evaluations and achieves state-of-the-art fine-tuned performance on datasets like NQ, HotpotQA, and OTT-QA.
The retrieval model is an indispensable component for real-world knowledge-intensive tasks, e.g., open-domain question answering (ODQA). As separate retrieval skills are annotated for different datasets, recent work focuses on customized methods, limiting the model transferability and scalability. In this work, we propose a modular retriever where individual modules correspond to key skills that can be reused across datasets. Our approach supports flexible skill configurations based on the target domain to boost performance. To mitigate task interference, we design a novel modularization parameterization inspired by sparse Transformer. We demonstrate that our model can benefit from self-supervised pretraining on Wikipedia and fine-tuning using multiple ODQA datasets, both in a multi-task fashion. Our approach outperforms recent self-supervised retrievers in zero-shot evaluations and achieves state-of-the-art fine-tuned retrieval performance on NQ, HotpotQA and OTT-QA.