Search-based Neural Structured Learning for Sequential Question Answering

By Mohit Iyyer, Wen-tau Yih, Ming-Wei Chang. 2017

278 citations recorded by the source corpus.

Recent work in semantic parsing for question answering has focused on long and complicated questions, many of which would seem unnatural if asked in a normal conversation between two humans. In an effort to explore a conversational QA setting, we present a more realistic task: answering sequences of simple but inter-related questions. We collect a dataset of 6,066 question sequences that inquire about semi-structured tables from Wikipedia, with 17,553 question-answer pairs in total. To solve this sequential question answering task, we propose a novel dynamic neural semantic parsing framework trained using a weakly supervised reward-guided search. Our model effectively leverages the sequential context to outperform state-of-the-art QA systems that are designed to answer highly complex questions.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Natural Language Processing

Citations in the corpus

CoQA: A Conversational Question Answering Challenge (2018, background): “n advantage of our setup is that the questioner can validate the answerer on the spot resulting in high agreement data. Conversational Modeling Our focus is on questions that appear in a conversation.Iyyer et al. (2017) andTalmor and Berant(2018) break down a complex question into a series of simple”
QuAC: Question Answering in Context (2018, methodology): “CoQA (Reddy et al., 2018) CSQA (Saha et al., 2018) CQA (Talmor and Berant, 2018) SQA (Iyyer et al., 2017) NarrativeQA (Kocisk´y et al., 2017) TriviaQA (Joshi et al., 2017) SQuAD 2.0 (Rajpurkar et al., 2018) MS Marco (Nguyen et al., 2016) NewsQA (Trischler et al., 2016) Table 1: Comparison of the QUA”
A Discrete Hard EM Approach for Weakly Supervised Question Answering (2019, methodology): “In reward-based methods, a reward function is defined as a prior, and the model parameters are updated with respect to it (Iyyer et al., 2017; Liang et al., 2017, 2018).”
UL2: Unifying Language Learning Paradigms (2022, background): “We use several component tasks from UnifiedSKG (Xie et al., 2022), namely WikiTQ (Pasupat & Liang, 2015), CompWQ (Talmor & Berant, 2018), FetaQA (Nan et al., 2021), HybridQA (Chen et al., 2020), WikiSQL (Zhong et al., 2017), TabFat (Chen et al., 2019), Feverous (Aly et al., 2021), SQA (Iyyer et al.,”
LongForm: Effective Instruction Tuning with Reverse Instructions (2023, neutral): “They cover a diverse range of topics, including events, careers, biographies, and writing tasks such as reviews, summaries and essays – and this is just looking at the 11% of our LongForm-C dataset with reverse instructions (RI).”
UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models (2022, background): “, 2019a) Database Dialog SQL SQA (Iyyer et al., 2017) Table Multi turn Answer”
Neural Approaches to Conversational AI (2018, neutral): “ns of entities. 27 Figure 3.6: The examples from two conversational KB-QA datasets. (Left) An example question sequence created from a compositional question intent in the SQA dataset. Figure credit: Iyyer et al. (2017). (Right) An example dialogue from the CSQA dataset. Figure credit: Saha et al. (”
TabFact: A Large-scale Dataset for Table-based Fact Verification (2019, background): “(Pasupat & Liang, 2015), Sequential Q&A (Iyyer et al., 2017), and WikiSQL (Zhong et al., 2017), for which approaches have been extended to handle large-scale resources like Wikipedia (Bhagavatula et al., 2013).”
Interpretation of Natural Language Rules in Conversational Machine Reading (2018, background): “In this space, previous approaches have been looking primarily at QA dialogs about images (Das et al., 2017) and knowledge graphs (Saha et al., 2018; Iyyer et al., 2017).”
JuICe: A Large Scale Distantly Supervised Dataset for Open Domain Context-based Code Generation (2019, background): “, 2018) ATIS (Zettlemoyer and Collins, 2009) SequentialQA (Iyyer et al., 2017) SCONE (Long et al.”
Coupling Retrieval and Meta-Learning for Context-Dependent Semantic Parsing (2019, background): “Recently, context-dependent semantic parsing has drawn plenty of attention (Long et al., 2016; Iyyer et al., 2017; Iyer et al., 2018; Suhr et al., 2018; Suhr and Artzi, 2018), where the generation of logical forms is conditioned on the context environment.”
Maximum Margin Reward Networks for Learning from Explicit and Implicit Supervision (2017, methodology): “In our companion paper, (Iyyer et al., 2017), we used a simple form of approximated reward to get feedback as early as possible during search.”
A dataset and baselines for sequential open-domain question answering (2018, methodology): “Both Iyyer et al. (2017) and Talmor and Berant (2018) answer complex questions by decomposing each into a sequence of simple questions.”
Robust Text-to-SQL Generation with Execution-Guided Decoding (2018, methodology): “A first generation of these approaches (e.g. by Iyyer, Yih, and Chang (2017)) has focused on adapting tools from (neural) machine translation, such as deep sequence-tosequence (seq2seq) architectures with attention and copying mechanisms.”
TAPEX: Table Pre-training via Learning a Neural SQL Executor (2021, methodology): “This improvement is also a surprise to us since SQA is a conversational dataset while our pre-training task is context-free.”
A Survey on Semantic Parsing (2018, background): “Iyyer et al. [2017] and Peng et al. [2017] use this type of framework. Misra et al. [2018] explain how these various training paradigms relate to each other and provide a generalized update equation making it easier to choose an approach that would best suit a particular problem.”