Date: Wednesday, 22 June 2022

Time: 5:00 p.m. - 6:00 p.m.

Via Zoom: https://hku.zoom.us/j/92530313632?pwd=ckhDejJvWDRTekZkaWdJZ1lJVXNwQT09

Meeting ID: 925 3031 3632

Password: 441117

5:00 p.m.

Presenter: Mr. Yunfan LI, PhD candidate
Primary Supervisor: Prof. Pengtao LIU
Presentation Title: Blastocyst-like structure to investigate early embryo development and animal cloning
Abstract: All mammals begin humbly from a bona fide totipotent cell, the zygote. Numerous cell cleavages cycles after fertilization rapidly induces a boost of developmental potency and subsequently form the blastocyst following the first-lineage segregation, which is comprised of the trophectoderm (the extra-embryonic part) and the inner cell mass (the embryonic part). Many efforts based on deep multi-omics analysis for investigating the underlying mechanisms of early embryo development have enlightened people to reconstruct blastocyst-like structures, namely blastoids in vitro from various mouse and human stem cell lines. With the well-established knowledge of first cell fate commitment, we expanded the developmental potential of mouse embryonic stem cells and generated mouse blastoids with simplified protocols. These blastoids are applied to investigate the molecular landscape of early embryonic development and the predicaments in generating animal clones. 

5:30 p.m.

Presenter: Mr. Shichao MA, PhD candidate
Primary Supervisor: Dr. Joshua HO
Presentation Title: Smartphone-based cardiac auscultation and disease screening through deep learning
Abstract: Advances in mobile technologies and the widespread use of smartphones provide opportunities to deliver healthcare services, including cardiac auscultation, over geographical barriers. We propose a method of acquiring high-quality phonocardiogram data through smartphones that uses a deep neural network to consistently detect heartbeats in real-time to help users locate the ideal auscultation position and automate the recording process. We further develop and validate a DWT-based approach to reduce ambient noises with low computational consumption. All algorithms and models are deployed as a lightweight smartphone application to facilitate data collection and cardiac auscultation that can be used in a telemedicine setting. Moreover, we implement a self-supervised learning algorithm to support the screening of cardiovascular diseases using smartphone-based phonocardiograms. This approach can reduce the need for well-labelled data while achieving high disease detection accuracy.  

ALL ARE WELCOME

Should you have any enquiries, please feel free to contact Miss Cynthia Cheung at 3917 9748.