A Conversational Recommender Skill for Conference Content

Abstract

Conversational information access systems aim to assist users in task-oriented exchanges of information in a conversational manner. Supporting multiple user goals, such as informational search, recommendation and exploratory information gathering. This thesis presents the implementation and development cycle of a conversational recommender skill for conference content, an instance of a conversational information access system. The project is a part of a larger one, with the final goal to produce a product to be available as a conversation conference assistant at the European Conference on Information Retrieval (ECIR) in April 2022. The skill features rich responses, making use of multi-modality in the form of an external screen.

Conversational information access systems aim to assist users in task-oriented exchanges of information in a conversational manner. Supporting multiple user goals, such as informational search, recommendation and exploratory information gathering. This thesis presents the implementation and development cycle of a conversational recommender skill for conference content, an instance of a conversational information access system. The project is a part of a larger one, with the final goal to produce a product to be available as a conversation conference assistant at the European Conference on Information Retrieval (ECIR) in April 2022. The skill features rich responses, making use of multi-modality in the form of an external screen.