AI solutions to real world problems
The potential of AI was on full display recently at the Alibaba Cloud Singapore AI Hackathon 2024. The projects that participants came up with were some of the most practicable, yet innovative solutions to real-world challenges that people face in everyday life.
Nicholas Koh, team lead of Amaris, third prize winner in the competition, said: “What excites me about AI is its potential to revolutionize virtually every aspect of human life, from healthcare to transportation, by enabling unprecedented levels of automation, efficiency, and innovation.”
?AI is exciting because it can automate boring tasks,” Jess Tang, the leader of second prize winner team Mavic said. “The potential of AI is to allow for quick generation of new ideas that can spark off new designs and experiments.”
?
1st prize: “BabyNoCry: An Innovative AI Baby Monitor”
The top prize of S$10,000 went to AliCloudHackers, who created an innovative AI baby monitor. The solution leveraged advanced sound, speech, and sentiment recognition technologies to providing real-time updates and insights to parents, allowing them to be able to still monitor their infants when they are some distance away.
It would be a welcome solution to a whining issue to many a new parent, who can perfectly emphathize on the need for it.
How it works
After performing some initial checks and sending its data to the cloud for analysis, it will input the data into a user-friendly dashboard, providing parents with a comprehensive overview of the child’s condition.
The solution utilizes the following Alibaba Cloud enabling technologies:
Joint 2nd prize: “Physical Threat Detection: Enhancing Security for Businesses”
A team from Hendricks Corp, a provider of AI and video solutions, won joint 2nd prize (S$3,000) with their solution, Physical Threat Detection: Enhancing Security for Businesses”. Their solution addressed an area of increasing concern – physical security for businesses.
Said their team lead, Souhail Meftah, a Senior Technical Product Manager with the company: “Based on our market-research, we noticed a glaring problem in the utilization of CCTV for security in Singapore. While the number of cameras and coverage is substantial, they are underutilized in the sense that, 85% of them are not monitored; thus, they act only as deterrent agents and post-event evidence and action can only be taken post-incident, and an operator typically needs to comb through hundreds of hours of footage from several cameras to trace the incident.”
“This drove us to develop a solution that not only automates all the manual work needed by the operators, but also enables the utilization of CCTVs for immediate intervention to active threats,” he added.
How it works
The solution utilizes a 3D-convolution based classifier to classify a person's action based on a sequence of frames of that person. Following the classifications, it provides real-time alerts when a person's behavior is detected as suspicious/threatening. Using image-to-text model (PromptCap) and then text-to-speech model (Intelligent Speech Interaction it is then sent to security personnel on-site for their action.
The following Alibaba Cloud enabling technologies were used:
?
Joint 2nd prize: “GenieSpark: AI-Powered Product Review Analyzer”
GenieSpark by Team Mavic, is an AI-powered product review analyzer that tells businesses which reviews are driving sales, and which reviews are losing sales for them. Genie Spark can deliver an analysis, by analyzing user behaviour when they are on a product page.
“Data analysis is always a challenge for e-commerce businesses because of lack of reviews and fake reviews,” team leader Jess Tang said. “We wanted a way to overcome these challenges and be able to help businesses identify real shopper behaviour, beyond the limitations of the data published onto the site. With AI, we are able to use self-learning to draw these correlations much faster and better,” Jess added.
领英推荐
Alibaba Cloud’s enabling technologies:
On the event, Jess said: We enjoyed being exposed to new technologies, like understanding how Alibaba Cloud works and the various LLMs available via this infrastructure. We also enjoyed meeting new people and hearing their perspectives and ideas on how to solve problems with AI.”
?
Joint 3rd prize: “Automated Marking Machine for Paper-Based Homework”
Team Cabbage’s solution would be a boon to teachers all over, saving them hours and hours of manual work. It is an automated paper homework correction machine that grades students' paper homework, imprinting them on the papers. The machine can grade over 100 papers in one minute with an accuracy rate exceeding 95%. The innovation frees teachers to do more meaningful work like preparing coursework and conducting classes.
Tony Lu Chenghao, leader of Team Cabbage, said: “Students want more explanations on why it is wrong, not just what is wrong. As teachers don’t have time to explain to students one by one, our solution will do the work.”
The project utilised the following Alibaba Cloud enabling technologies:
“More people can now participate in developing AI-native applications,” Tony said. The democratization of AI technology means that it is not just limited to tech giants and specialists; it's becoming a tool that anyone can use to create, innovate, and solve problems. The potential here lies in the variety and depth of applications that can emerge from such a broad base of users, driving innovation in ways that were previously unimaginable.
?
Joint 3rd prize: “DormChecker: Enhancing Living Conditions in Dormitories with AI Technology”
Team Amaris’s project/solution looked into the wellbeing of the migrant workers. Their project was an Android application that takes in an image of a room or a dormitory bunk and using multimodal large language models, provides a rating for hygiene and cleanliness. The platform also provides an AI chatbot for users to ask question, with answers derived from contents relating to official published guidelines by Singapore’s Ministry of Manpower.
“We also strongly believe in the vision of the project, which is to improve the living conditions of the dormitory workers,” Nicholas Koh, team lead of Amaris said.
“I think it was a good challenge and very good exposure to the Aliyun cloud services,” he added.
Alibaba Cloud enabling technologies:
?
Joint 3rd prize: AI Xin: Streamlining Workflows with Robotic Process Automation
Team AI Xin’s project aims to streamline user workload through the implementation of robotic process automation. By providing a swift and efficient method to review and provide iterative feedback on the performed steps, the project significantly reduces routine tasks, allowing employees to focus on higher-value responsibilities, and increasing overall productivity.
Leveraging Qwen-VLmax for LVLM visual inference to verify the graphical user interface (GUI) against recorded data enables more robust GUI robotic process automation (RPA).
Alibaba Cloud enabling technologies: