Lina Yao, University of New South Wales

Data-Efficient Visual Analysis with Limited Supervision

Abstract

It is difficult to collect and generate sufficient annotated data in the real world due to limited resources, such as costly human labours for annotation, security and privacy concerns. Therefore, it is urgent to develop methods to complete the visual analysis mission with only limited supervision. In this talk, I will share my team’s recent research progress about data-efficient visual analysis on various tasks in terms of visual classification, action detection, and event detection.

Bio

Lina Yao is an Associate Professor with the School of Computer Science and Engineering, UNSW. Her research interest lies in developing generalizable and explainable data-efficient machine learning methods—as well as designing systems and interfaces—to enable novel ways of human-machine interactions, including an improved understanding of challenges such as robustness, trust, explainability, and resilience of human-autonomy partnership. She is serving as Associate Editor for ACM Transactions on Sensor Networks, Knowledge-based Systems, and Frontiers in Big Data in Recommender Systems section. More details can be found at https://www.linayao.com/.