Improving the Community Question Retrieval Performance Using Attention-Based Siamese LSTM

Abstract

In this paper, we focus on the problem of question retrieval in community Question Answering (cQA) which aims to retrieve from the community archives the previous questions that are semantically equivalent to the new queries. The major challenges in this crucial task are the shortness of the questions as well as the word mismatch problem as users can formulate the same query using different wording. While numerous attempts have been made to address this problem, most existing methods relied on supervised models which significantly depend on large training data sets and manual feature engineering. Such methods are mostly constrained by their specificities that put aside the word order and ignore syntactic and semantic relationships. In this work, we rely on Neural Networks (NNs) which can learn rich dense representations of text data and enable the prediction of the textual similarity between the community questions. We propose a deep learning approach based on a Siamese architecture with LSTM networks, augmented with an attention mechanism. We test different similarity measures to predict the semantic similarity between the community questions. Experiments conducted on real cQA data sets in English and Arabic show that the performance of question retrieval is improved as compared to other competitive methods.

Arabic Medical Community Question Answering Using ON-LSTM and CNN

“…This problem was solved by using attention techniques such as multi-head attention that forces the model to focus on the important parts of the sentence [22]. In this context, [31] proposed a recent approach to Arabic QA that used a Siamese LSTM networks based model with an attention mechanism. They evaluated their model on the cQA dataset in English and Arabic using different similarity measures and reported a performance improvement compared with the other competitive approaches.…”

“…• In some social media content (such as medical forums) English terms are used which result in mixed language text content (Arabic and English) [29]. Most of the current studies on Arabic cQA use the similarity between the user question and the archived questions in the collection or corpus (question-question similarity) [7,20,31,35,39]. However, questions tend to be short so only considering the question-question similarity results means there is sparse information to work with.…”

Technological troubleshooting based on sentence embedding with deep transformers

“…Human annotators are employed to evaluate and find 2 to 30 relevant solutions for each query, resulting in a high management cost. This QR approach is extended in (Othman et al, 2020) by enhancing the neural network architecture with an attention mechanism to determine which words in the questions should receive more attention during the embedding phase. Again, the approach requires a large number of manually labeled pairs of queries and relevant sentences, about 30% more than the previous study.…”