R³: Reinforced Ranker-Reader for Open-Domain Question Answering

Abstract

In recent years researchers have achieved considerable success applying neural network methods to question answering (QA). These approaches have achieved state of the art results in simplified closed-domain settings such as the SQuAD (Rajpurkar et al. 2016) dataset, which provides a pre-selected passage, from which the answer to a given question may be extracted. More recently, researchers have begun to tackle open-domain QA, in which the model is given a question and access to a large corpus (e.g., wikipedia) instead of a pre-selected passage (Chen et al. 2017a). This setting is more complex as it requires large-scale search for relevant passages by an information retrieval component, combined with a reading comprehension model that “reads” the passages to generate an answer to the question. Performance in this setting lags well behind closed-domain performance. In this paper, we present a novel open-domain QA system called Reinforced Ranker-Reader (R3), based on two algorithmic innovations. First, we propose a new pipeline for open-domain QA with a Ranker component, which learns to rank retrieved passages in terms of likelihood of extracting the ground-truth answer to a given question. Second, we propose a novel method that jointly trains the Ranker along with an answer-extraction Reader model, based on reinforcement learning. We report extensive experimental results showing that our method significantly improves on the state of the art for multiple open-domain QA datasets.

Weakly-Supervised Visual-Retriever-Reader for Knowledge-based Question Answering

“…The advanced reading comprehension model (Chen et al, 2017a;Banerjee et al, 2019) split this complex task into two steps: a retriever selects some most relevant documents from a corpus to a question, and a reader produces answer according to the documents from retriever. Some previous work (Kratzwald and Feuerriegel, 2018;Lee et al, 2018;Das et al, 2019;Wang et al, 2018) train the end-to-end models to rerank in a closed set. Although these mod-els are better at retrieval, they can hardly scale to larger corpora.…”

TopiOCQA: Open-domain Conversational Question Answering with Topic Switching

“…Models for open-domain QA often follow a two-stage process: (1) A retriever selects a small collection of documents relevant to the question from a big corpus (e.g., Wikipedia), (2) a reader extracts or generates an answer from the selected documents. While classical approaches rely on counting-based bag-of-words representations like TF-IDF or BM25 (Chen et al, 2017;Wang et al, 2018;Yang et al, 2019), more recent deep learning approaches learn dense representations of the questions and document through a dual-encoder framework Karpukhin et al, 2020). In such learned retriever setups, document retrieval is done efficiently using Maximum Inner Product Search (MIPS, Shrivastava and Li, 2014).…”

Extract-Select: A Span Selection Framework for Nested Named Entity Recognition with Generative Adversarial Training

“…1) Candidate set representation: To represent the entity span candidate set C, we propose to encode the information of both entity type y and sequence X into p C . Specifically, we build entity-aware sequence representation with Match-LSTM (Wang et al, 2018b), by matching the entity marker up with the sequence:…”