A Compare-Aggregate Model with Latent Clustering for Answer Selection

First, we explore the effect of additional directions for obtaining the best performance in the answerselection information by adopting a pretrained language model to task. First, we explore the effect of additional information compute the vector representation of the input text and by applying by adopting a pretrained language model (LM) to compute the vector transfer learning from a large-scale corpus. Second, we enhance representation of the input text. Recent studies have shown that the compare-aggregate model by proposing a novel latent clustering replacing the word-embedding layer with a pretrained language method to compute additional information within the target model helps the model capture the contextual meaning of words corpus and by changing the objective function from listwise in the sentence [2, 6]. Following this study, we select an ELMo [6] to pointwise. To evaluate the performance of the proposed approaches, language model for this study. Furthermore, we investigate the experiments are performed with the WikiQA and TREC-applicability of transfer learning (TL) by using a large-scale corpus QA datasets. The empirical results demonstrate the superiority of that is created for relevant-sentence-selection task (i.e., questionanswering our proposed approach, which achieve state-of-the-art performance