Abstract
Nowadays information systems have been increasingly used in assisting our everyday decision-makings, from recommending movies to watch, products to purchase, to even providing treatment to patients. Such systems enable us to have every kind of data we possibly want at our fingertips. on the contrary, information system become more involved in making crucial decisions affecting human livelihoods, e.g., clinical decision-making criminal's incarceration. It brings up serious concerns on whether the systems could provide reliable service, and urges us to improve the reliability of the information systems.
The reliability of the decision-making can be evaluated in various aspects. First, accuracy undoubtedly is the foundation of a reliable information system. Users' trust on the system will be hurt if they are presented with instances that are subsequently found to be inferior. Besides, only returning the most relevance results to users is insufficient to help users to perceive the value of the provided information. To gain users' trust, the system also needs to be more transparent and make it more explicit why the users should pay attention to those returned results. In addition, served as the inter-media between information provider and consumer, modern information systems need to be trustful for both the consumer and the information provider. For the candidate instances, a reliable decision-making should not create discriminatory or unjust impacts when comparing across different demographics.
In this thesis, we focus on three aspects of the reliability of the decision-making in information system, accuracy, fairness, and transparency,. More specifically, we present three main tasks to enhance the modern information systems: 1) efficient online learning to rank; 2) fair online learning to rank; and 3) explainable recommendation. Our study provides a deep and thorough understanding of the importance of the reliability in information systems, and improves the reliability of the system by providing more effective, fair and transparent service.
Rigorous theoretical analysis and extensive empirical evaluation validated the approaches' applicability in various contexts and applications.
