Discriminative Information Retrieval for Knowledge Discovery
This work addresses the initial step in text-based question answering, offering a domain-specific improvement for knowledge discovery tasks.
The paper tackled the problem of improving recall in answer candidate passage retrieval for question answering by proposing a discriminative information retrieval framework based on linguistic features, resulting in a 44% improvement in recall for candidate triage.
We propose a framework for discriminative Information Retrieval (IR) atop linguistic features, trained to improve the recall of tasks such as answer candidate passage retrieval, the initial step in text-based Question Answering (QA). We formalize this as an instance of linear feature-based IR (Metzler and Croft, 2007), illustrating how a variety of knowledge discovery tasks are captured under this approach, leading to a 44% improvement in recall for candidate triage for QA.