Zhi Qin Tan
About
My research project
Robust multi-modal learning for data-driven batch and sequential inferenceA wide range of applications, such as computer vision, target tracking, finance, navigation, and robotics, involve inherently multi-modal, high-dimensional data (such as images and texts). This wealth of data provides an exciting opportunity to develop artificial intelligence (AI) techniques and also present a major challenge to develop AI solutions that generalise well when deployed in real world due to the curse of dimensionality, data heterogeneity and bias.
This project will develop robust machine learning techniques that capture and express the complexity of multimodal data in real-world AI applications, using digital health as a clinically important and methodologically convenient exemplar. Our research vision is to establish a principled process to incorporate information and uncertainty present in multi-modal, high-dimensional digital health data to improve robustness of AI solutions and accelerate their adoption. The project will focus on both batch learning and sequential, online learning.
Supervisors
A wide range of applications, such as computer vision, target tracking, finance, navigation, and robotics, involve inherently multi-modal, high-dimensional data (such as images and texts). This wealth of data provides an exciting opportunity to develop artificial intelligence (AI) techniques and also present a major challenge to develop AI solutions that generalise well when deployed in real world due to the curse of dimensionality, data heterogeneity and bias.
This project will develop robust machine learning techniques that capture and express the complexity of multimodal data in real-world AI applications, using digital health as a clinically important and methodologically convenient exemplar. Our research vision is to establish a principled process to incorporate information and uncertainty present in multi-modal, high-dimensional digital health data to improve robustness of AI solutions and accelerate their adoption. The project will focus on both batch learning and sequential, online learning.
Publications
Acquiring fine-grained object detection annotations in unconstrained images is time-consuming, expensive, and prone to noise, especially in crowdsourcing scenarios. Most prior object detection methods assume accurate annotations; A few recent works have studied object detection with noisy crowdsourced annotations, with evaluation on distinct synthetic crowdsourced datasets of varying setups under artificial assumptions. To address these algorithmic limitations and evaluation inconsistency, we first propose a novel Bayesian Detector Combination (BDC) framework to more effectively train object detectors with noisy crowdsourced annotations, with the unique ability of automatically inferring the annotators' label qualities. Unlike previous approaches, BDC is model-agnostic, requires no prior knowledge of the annotators' skill level, and seamlessly integrates with existing object detection models. Due to the scarcity of real-world crowdsourced datasets, we introduce large synthetic datasets by simulating varying crowdsourcing scenarios. This allows consistent evaluation of different models at scale. Extensive experiments on both real and synthetic crowdsourced datasets show that BDC outperforms existing state-of-the-art methods, demonstrating its superiority in leveraging crowdsourced data for object detection.