MULTIDIMENSIONAL RELEVANCE IN TASK-SPECIFIC RETRIEVAL

Relevance is the core notion in Information Retrieval. Several criteria of relevance have been proposed in the literature. Relevance criteria are strongly related to the search task. Thus, it is important to employ the criteria that are useful for the considered search task. This research explores the concept of multidimensional relevance in a specific search-task. In the first phase of this PhD thesis, we aim to investigate the relationship between the search tasks and the considered relevance dimensions. We performed an exploratory study on different search tasks in the Microblog search context, and we identify some related relevance dimensions. Our findings show that there is a relation between a task and specific relevance dimensions. This suggests that in different search-tasks, some relevance dimensions should be prioritized while others should not be considered. In the second part, we propose an approach that can be used to combine more than one relevance dimension. In particular, given that recent advancements in deep neural networks enable several learning tasks to be solved simultaneously, we examine the possibility of modeling multidimensional relevance by jointly solving a retrieval task, to learn topical relevance, and a classification task, to learn additional relevance dimensions. To instantiate and evaluate the proposed model, we consider three query-independent relevance dimensions beyond topicality, i.e., readability, trustworthiness, and credibility. The findings show that the proposed joint modeling can improve the performance of the retrieval task.