Open Domain Question Answering Using Early Fusion of Knowledge Bases and Text

Abstract

Open Domain Question Answering (QA) is evolving from complex pipelined systems to end-to-end deep neural networks. Specialized neural models have been developed for extracting answers from either text alone or Knowledge Bases (KBs) alone. In this paper we look at a more practical setting, namely QA over the combination of a KB and entitylinked text, which is appropriate when an incomplete KB is available with a large text corpus. Building on recent advances in graph representation learning we propose a novel model, GRAFT-Net, for extracting answers from a question-specific subgraph containing text and KB entities and relations. We construct a suite of benchmark tasks for this problem, varying the difficulty of questions, the amount of training data, and KB completeness. We show that GRAFT-Net is competitive with the state-of-the-art when tested using either KBs or text alone, and vastly outperforms existing methods in the combined setting.

Look before you Hop: Conversational Question Answering over Knowledge Graphs Using Judicious Context Expansion

“…While templates continued to be a strong line of work due to its focus on interpretability and generalizability [1,2,4,7,35], a parallel thread has focused on neural methods driven by performance gains [15,20,31]. Newer trends include shifts towards more complex questions [19,21,34], and fusion of knowledge graphs and text [31,33]. However, none of these approaches can deal with incomplete questions in a conversational setting.…”

PullNet: Open Domain Question Answering with Iterative Retrieval on Knowledge Bases and Text

“…For 1-hop questions in MetaQA (which is identical to WikiMovies), our model is compara-ble to the state-of-the-art 6 . For the other three settings, the performance of our re-implementation is slightly worse than the results reported in by original GRAFT-Net paper (Sun et al, 2018); this is likely because we use a simpler retrieval module.…”

“…GRAFT-Net (Sun et al, 2018) supports multihop reasoning on both KBs and text by introducing a question subgraph built with facts and text, and uses a learned graph representation (Kipf and Welling, 2016;Li et al, 2016;Schlichtkrull et al, 2017;Scarselli et al, 2009) to perform the "reasoning" required to select the answer. We use the same representation and reasoning scheme as GRAFT-Net, but do not require that the entire graph be retrieved in a single step.…”

“…In this paper we follow previous research (Sawant et al, 2019;Sun et al, 2018) in deriving answers using both a corpus and a KB. We focus on tasks in which questions require compositional (sometimes called "multi-hop") reasoning, and a setting in which the KB is incomplete, and hence must be supplemented with information extracted from text.…”

Discourse-Aware Semantic Self-Attention for Narrative Reading Comprehension

“…In this work we focus on the recent Narra-tiveQA (Kocisky et al, 2018) dataset that was designed not to be easy to answer and that requires a model to read narrative stories and answer questions about them. In terms of model architecture, previous work in reading comprehension and question answer- ing has focused on integrating external knowledge (linguistic and/or knowledge-based) into recurrent neural network models using Graph Neural Networks (Song et al, 2018), Graph Convolutional Networks (Sun et al, 2018;De Cao et al, 2019), attention (Das et al, 2017;Mihaylov and Frank, 2018;Bauer et al, 2018) or pointers to coreferent mentions (Dhingra et al, 2017).…”