Applying deep learning to answer selection: A study and an open task

By Minwei Feng, Bing Xiang, Michael R. Glass, Lidan Wang, Bowen Zhou. 2015

390 citations recorded by the source corpus.

We apply a general deep learning framework to address the non-factoid question answering task. Our approach does not rely on any linguistic tools and can be applied to different languages or domains. Various architectures are presented and compared. We create and release a QA corpus and setup a new QA task in the insurance domain. Experimental results demonstrate superior performance compared to the baseline methods and various technologies give further improvements. For this highly challenging task, the top-1 accuracy can reach up to 65.3% on a test set, which indicates a great potential for practical use.

Subfields: Artificial Intelligence, 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 Natural Language Processing

Citations in the corpus

A Structured Self-attentive Sentence Embedding (2017, background): “…networks (Kalchbren-ner et al., 2014; dos Santos & Gatti, 2014; Kim, 2014) as an intermediate step in creating sentence representations to solve a wide variety of tasks including classiﬁcation and ranking (Yin & Sch¨utze, 2015; Palangi et al., 2016; Tan et al., 2016; Feng et al., 2015).”
ABCNN: Attention-Based Convolutional Neural Network for Modeling Sentence Pairs (2015, background): “How to model a pair of sentences is a critical issue in many NLP tasks such as answer selection (AS) (Yu et al., 2014; Feng et al., 2015), paraphrase identification (PI) (Madnani et al., 2012; Yin and Schütze, 2015a), textual entailment (TE) (Marelli et al., 2014a; Bowman et al., 2015a) etc.”
Attentive Pooling Networks (2016, methodology): “For the InsuranceQA dataset, we use the 100-dimensional vectors that were trained by Feng et al. (2015) using word2vec (Mikolov et al., 2013).”
Semi-supervised Question Retrieval with Gated Convolutions (2015, background): “Recent work on answer selection on community QA forums, similar to our task of question retrieval, has also involved the use of neural network architectures (Severyn and Moschitti, 2015; Wang and Nyberg, 2015; Shen et al., 2015; Feng et al., 2015; Tan et al., 2015).”
Inner Attention based Recurrent Neural Networks for Answer Selection (2016, methodology): “We adopt the Geometric mean of Euclidean and Sigmoid Dot (GESD) proposed in (Feng et al., 2015) to measure the similarity be-”
Analysis of Wikipedia-based Corpora for Question Answering (2018, background): “Feng et al. (2015) presented INSURANCEQA comprising 16K+ questions on insurance contexts.”
SelQA: A New Benchmark for Selection-Based Question Answering (2016, methodology): “Several corpora have been created for these tasks [1, 2, 3], allowing researchers to build effective question answering systems [4, 5, 6] with the aim of improving reading comprehension through understanding and reasoning of natural language.”
Polisis: Automated Analysis and Presentation of Privacy Policies Using Deep Learning (2018, background): “Over the past few years, deep learning has yielded superior results to traditional retrieval techniques in this domain [51, 52, 66].”
MultiCQA: Zero-Shot Transfer of Self-Supervised Text Matching Models on a Massive Scale (2020, background): “Zero-shot transfer of text matching models is particularly challenging in setups of non-factoid answer selection (Cohen et al., 2018; Tay et al., 2017; Feng et al., 2015; Verberne et al., 2010) and question similarity (Nakov et al., 2017; Lei et al., 2016).”
Beyond [CLS] through Ranking by Generation (2020, background): “Considerable body of work exists on the use of NNs for this task (Feng et al., 2015; Severyn and Moschitti, 2015; Tan et al., 2016; dos Santos et al., 2016; Rao et al., 2016; Wang et al., 2017), where the most recent ones use BERT-based models that perform discrimination based on the special [ CLS…”
Advances and Challenges in Conversational Recommender Systems: A Survey (2021, methodology): “Then, a matching function is used to combine the two representations and output a matching probability [63, 163, 138, 45, 174].”
Named Entity Recognition in the Medical Domain with Constrained CRF Models (2017, methodology): “For the binary classifier, we make use of IBM’s Natural Language Classifier service3 which is relies on a Convolutional Neural Network combined with word embeddings (Feng et al., 2015).”
Improved Representation Learning for Question Answer Matching (2016, methodology): “Compared to Architecture II in (Feng et al., 2015), which involved a large number of CNN filters, (N) model also has fewer parameters.”
IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models (2017, background): “InsuranceQA [10] is one of the most studied question-answering dataset.”
Denoising Bodies to Titles: Retrieving Similar Questions with Recurrent Convolutional Models (2015, background): “For instance, Feng et al. (2015) explore different similarity measures beyond cosine similarity, and Tan et al. (2015) adopt the neural attention mechanism over RNNs to generate better answer representations given the questions as context.”
Learning to Answer Questions from Wikipedia Infoboxes (2016, background): “Models like (Feng et al., 2015) and (Wang and Ny-berg, 2015) generate feature representations of questions and answers using neural networks, computing the similarity of these representations to select an answer.”
emrQA: A Large Corpus for Question Answering on Electronic Medical Records (2018, background): “…such as MCTest (Richardson et al., 2013), biological process modeling (Berant et al., 2014), BioASQ (Tsatsaronis et al., 2015), InsuranceQA (Feng et al., 2015), etc have been limited in scale (500-10K) because of the complexity of the task or the need for expert annotations that cannot be…”
LSDSem 2017: Exploring Data Generation Methods for the Story Cloze Test (2017, methodology): “To train models for this task, Feng et al. (2015) have extended a corpus of question-answer pairs by automatically adding negative answers to each pair.”