Smart Campus Training Programme — Innovation & Technology Enrichment Course (Secondary Students) Day One

INSPIRE instructors were invited to deliver the Innovation & Technology Enrichment Course for secondary students as part of the Smart Campus Training Programme. The course was held on 20 April 2024 at Google Hong Kong's office in Times Square, Causeway Bay, from 09:30 to 14:15. Day One was divided into four parts: drone programming — automated flight mission challenge, drone 3D terrain scanning applications, AI experience — how to create an AI tutor, and AI experience — how to create a customer service AI.

Drone Programming — Automated Flight Mission Challenge

INSPIRE first introduced the components and STEM-related flight principles of the Tello Talent mini drone. After reminding students of drone safety and hidden risks, the instructor covered flight techniques including precise control, blind-spot awareness, and hand-eye coordination. Students then used the Tello App to practise controlling the drone in different directions. Next, they used Tello EDU to program flight commands, connecting devices to the Tello Talent via WiFi for 'one-to-one' and even 'one-to-many' flight modes to complete various flight missions.

Drone 3D Terrain Scanning Applications

For the 3D terrain scanning section, the instructor introduced aerial photogrammetry, its significance and evolving applications, and finally practical uses of 3D surveying within the school campus. Aerial photogrammetry is the science of making measurements from photographs to recover the exact positions of surface points. Students also had the opportunity to try mobile scanning apps, supporting computer-based calculations and modelling.

AI Experience — How to Create an AI Tutor

The AI section began with how to ask AI precise questions to obtain ideal answers, introducing Large Language Models (LLMs). LLMs are divided into base models, which predict the next word based on training data, and fine-tuned models, which are enhanced through human feedback to provide more considered responses. Three prompting techniques were covered: expressing questions clearly and specifically using delimiters; structured output to guide the AI step by step; and providing few-shot examples to steer the AI in the desired direction.

AI Experience — How to Create a Customer Service AI

In the final section, the instructor taught students how to create a bot on Poe, leveraging AI deep learning to build a customer service assistant for their own company or product. Students could upload documents or PDFs and break complex problems into small steps for the AI to handle. They used their own devices to try building bots on the spot, putting their learning into practice.