Answering Topical Information Needs Using Neural Entity-Oriented Information Retrieval and Extraction

In the modern world, search engines are an integral part of human lives. The field Information Retrieval (IR) is concerned with finding material (usually documents) unstructured nature text) that satisfies information need (query) from within large collections stored on computers). engine then displays a ranked list results relevant to our query. Traditional document retrieval algorithms match query using overlap words in both. However, last decade has seen focus shifting leveraging rich semantic available form entities. Entities uniquely identifiable objects or things such as places, events, diseases, etc. exist real fictional world. Entity-oriented systems leverage associated entities (e.g., names, types, etc.) better documents queries. Web would provide if they understand meaning This dissertation advances state-of-the-art IR by developing novel text (query, document, question, sentence, at level. To this end, aims fine-grained context which have been mentioned, for example, "oysters" food versus ecosystems. Further, automatically learn (vector) representations incorporate knowledge and about refines automatic understanding passages deep learning, artificial intelligence paradigm. utilizes extracted retrieve materials (text entities) interplay between studied addressing three related prediction problems: (1) Identify query, (2) Understand entity's (3) elaborate connection entity. research presented may be integrated into larger system designed answering complex topical queries dark chocolate health benefits require connections material, thus transforming engine. Awarded by: University New Hampshire, Durham, USA 1 September 2022. Supervised Laura Dietz. Available at: https://scholars.unh.edu/dissertation/2714/.