20 NOV 2023

Champions of the Hackathon at the 7th HackADay

a team comprising of three BSc Computer Science students FOR Lek Shyuen, WONG Yu Fai and ZHENISHBEK UULU Talantbek together with BEng Computer and Data Engineering student HARIYANTO Vanessa Laurel, under the coordination of Prof HANCKE Gerhard Petrus, was announced as the Champions of the Hackathon at the 7th HackADay event hosted by PwC Hong Kong. 

This year’s Hackathon, focusing on the theme of Securing AI, was held in October and open to undergraduate students from universities in Hong Kong and Macau.  Teams were tasked to design and implement an interactive application powered by artificial intelligence (AI) / machine learning (ML) with cybersecurity consideration built-in, over a span of two weeks.  Shortlisted teams then showcased their systems to a judging panel consisting of members from the PwC Hong Kong Cybersecurity and Cloud team together with representatives from Amazon Web Services and Microsoft Azure.  The prize presentation for the winners took place at the sky100 Hong Kong Observation Deck on 7 November 2023.   

03 NOV 2023

Second Place Winner in the 2023 ACM/IEEE TinyML Design Contest

Three PhD students, Lianming Huang, Yu Mao and Shangyu Wu, under the supervision of Professor Chun Jason Xue and Professor Nan Guan, in collaboration with three advisors from overseas universities including Dr Yufei Cui, a CS PhD graduate, have secured the Second Place Winner in the 2023 ACM/IEEE TinyML Design Contest

Their team, HugeRabbit, was given the challenge to design and implement an operational, open-source machine learning or deep learning algorithm that is capable of distinguishing life-threatening ventricular arrhythmias (i.e., irregularities in heartbeat) from intracardiac electrograms (IEGM) recordings and is deployable on a given microcontroller platform.  Among over 80 teams, HugeRabbit’s performance in detection precision, memory occupation, and inference latency earned them the second place.  The team was invited to present the solution and to receive a cash prize of US$1,000 in the 2023 International Conference on Computer-Aided Design (ICCAD) held in San Francisco, USA from 29 October to 2 November 2023. 

12 OCT 2023

Champion in the J.P.Morgan’s Hackathons: Code for Good 2023

A team comprising Third Year BSc Computer Science students Eklavya Agarwal and Vannes Wijaya and Third Year BEng in Computer and Data Engineering student Abhinav Balasubramanian, alongside with three students from other local universities, emerged as the Champion and each was awarded an iPad as the prize at the J.P.Morgan’s Hackathons: Code for Good 2023 held on 6 October.    

Code for Good is an annual hackathon challenge predominantly participated by undergraduates majoring in Computer Science and Engineering.  Throughout the 12-hour event, the team, in collaboration with mentors from J.P.Morgan, was given the task to build a web/mobile app aimed at enhancing financial education among Hong Kong students.  The team developed a web app featuring elements of gamification, exercises, videos, ranking board and data visualization.  In particular, their game's user interface (UI), user experience (UX) and sound features caught the panel’s attention, playing a vital role in contributing to the championship victory. 

04 SEP 2023

First Runner-up in the Final Year Project Competition of the IEEE (Hong Kong) Computational Intelligence Chapter

Ms Ruozhen He, an outstanding 2023 BSc Computer Science graduate who participated in the Research Mentorship Scheme under the supervision of Prof Rynson Lau, has won the First Runner-up in the Final Year Project Competition of the IEEE (Hong Kong) Computational Intelligence Chapter on 26 August 2023.   

In her accredited project titled Weakly-Supervised Camouflaged Object Detection with Scribble Annotations, Ms He together with other CityU collaborators proposed the first weakly-supervised method for camouflaged object detection using just scribble annotations as supervision.  This new method, which takes merely ∼10 seconds to label one image, is a significant improvement over the existing methods which cost ~60 minutes.  They further proposed a novel neural network to localize the boundaries and detect the camouflaged objects.  The mentioned paper, along with another collaboration by Ms He, titled Efficient Mirror Detection via Multi-Level Heterogeneous Learning, were both accepted as Oral Papers for presentation in the 37th Association for the Advancement of Artificial Intelligence (“AAAI”) Conference held in February 2023.