Extended Abstract: Building Preference Aware Trustworthy AI Through Fairness and Interpretability

Abstract

While AI has brought transformative changes to society, the opacity of current state-of-the-art models still raises ethical concerns in high-risk domains. It is necessary to improve the trustworthiness of these black-box models to make them more responsible. This paper presents five completed works on enhancing the trustworthiness of both traditional ML models and large language models (LLMs) from the perspectives of fairness and interpretability. Future work will extend the discussion of trustworthiness to multi-modal scenarios.