![]() Most of the questions and answers in these forums are often lengthy, informal, and contain abbreviations and grammatical mistakes. 1 As shown in this table, in CQA, each question has at least three parts: (i) question category, which is the category that the question belongs to (ii) question subject, which summarizes the question, and (iii) question body, which describes the question in details, and might contain useless or noisy parts as well. Table 1 shows two examples of questions, each with four answers, taken from the SemEval 2015 dataset. It should be noted that a good answer is an answer semantically relevant to the question, not necessarily the correct answer. As defined in SemEval 2015, in this task, the goal is to classify the answers given a question into three categories: (i) good, which are the answers that address the question well (ii) potentially useful to the user (e.g., because they can help educate him/her on the subject) (iii) bad or useless. In this paper, we address the task of answer selection. Thus, it is essential to automatically identify the best answers for each question. It means one has to go through all possible answers for assessing them, which is exhausting and time-consuming. Unfortunately, there is often no evaluation of the given answers in how much they are related to the question. In these forums, anyone can ask any question, and a question is answered by one or more members. Many applications such as sentiment analysis, recommender systems, relation extraction, and question answering integrate the information in KGs by linking the entities mentioned in the text to entities in the KGs.Ĭommunity question answering (CQA) forums, such as Stack Overflow and Yahoo! Answer provide new opportunities for users to share knowledge. They consist of entities and relationships among them. Knowledge graphs (KGs), such as DBpedia and BabelNet, are multi-relational graphs. The experimental results on three widely used datasets demonstrate that our proposed method is effective and outperforms the existing baselines significantly. Moreover, the model uses variational autoencoders (VAE) in a multi-task learning process with a classifier to produce class-specific representations for answers. It also uses the question category for producing context-aware representations for questions and answers. We also learn a latent-variable model for learning the representations of the question and answer, jointly optimizing generative and discriminative objectives. In this paper, we propose a novel answer selection method in CQA by using the knowledge embedded in KGs. Despite the obvious usefulness of commonsense and factual information in the KGs, to the best of our knowledge, KGs have been rarely integrated into the task of answer selection in community question answering (CQA). Training Set The whole manually annotated dataset is available at the dataset page.With the increasing popularity of knowledge graph (KG), many applications such as sentiment analysis, trend prediction, and question answering use KG for better performance. IMAGACT English verbs related to the linked scenesīabelNet English verbs related to the linked BS General data about the algorithm used, the training set and the amount of linked entities are reported below.ĭata about linking IMAGACT scenes linked to BS For simplicity verb search is available in two languages only: Italian and English. Scenes and BabelSynsets are represented by their ID and they can be visualized by clicking the Viceversa, to search a BabelSynset through a list of BabelNet verbs and see all the scenes linked to it. The two links above allow to explore the results of the linking procedure: it's possible to search an IMAGACT scene through a list of IMAGACT verbs and see all the BabelSynset that are linked to it or, This page reports the results of the IMAGACT-BabelNet linking algorithm. Search Scene (through IM verbs) Search BabelSynset (through BN verbs)
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |