Seminar Recording
This talk introduces Manifold Learning, the technique to uncover the intrinsic shape of the original data. We also discuss how different manifold learning paradigms can be designed to be incorporated to a dimensionality reduction technique to learn the accurate low-dimensional data representation.
About the presenter
Dr. Khanh Luong finished her PhD in Computer Science specializing in Data Science from Queensland University of Technology in 2019. She is currently working on the Applied Data science project at QUT Centre for Data Science.
Her research is concerned with dealing multiple aspect data, the data that is represented by multiple types of features or modalities or collected from multiple sources.
Techniques have been used and developed including dimensionality reduction techniques such as matrix factorization, subspace learning, deep learning auto-encoder.
The purpose is to correctly project multi-aspect data to a consensus low-dimensional representation before conducting any subsequent learning tasks such as clustering, classification, anomaly detection or topic modelling for text data.
Details:
| Location: | Online |
| Start Date: | 29/07/2021 [add to calendar] |
| Start Time: | 2pm |
| End Date: | 29/07/2021 |
| End Time: | 3pm (AEST) |
| Register: |