Typically completed by 10-14 year olds, students work collaboratively on a five hour project or challenge in self-managed groups. During the project, they use a CREST Discovery passport to record and reflect on their work. Afterwards, students communicate their findings as a group presentation.
Each pack provides teaching guides, kit lists, example timetables and suggested starter activities to help you run your day. Find out more about CREST Discovery Awards.
Idea development Research and develop your concept Start by researching your ideas. Look into what similar products already exist and how they work. • Where do they get their data from? What limitations are there? • What output does your tool produce? • Does your tool use supervised or unsupervised machine learning? • Do you teach your tool using training data? Where does this data come from? • How does your tool assess its performance? • What are the ethical considerations of your product. What is the worst thing that could happen? • How will your tool continue to improve once the product is on the market? • Where will you source live data from and how will this be used to provide feedback for the algorithm (set of rules that your machine follows)? Design your product You will need to make decisions about the physical designs of your product. Think about whether your idea could be integrated into a tool already in use, or whether it would be an entirely new product. • Would it be an app on a smartphone, or part of a fridge or a washing machine, or would it be something totally separate? • Who is your product aimed at? How will it be useful for them? • What is the physical format of your product? • What will you do to ensure that your tool will be safe and that people trust it? • How will you make your design both fit for purpose and attractive to consumers? • How is the machine learning integrated? Will your product collect data? • What is the reasoning behind your design? Marketing Have a look at what already exists, think about what makes your product unique. • Decide on a target audience • How will you reach them? • How will you convince them they need this product? • Write a 2 minute elevator pitch to explain and sell the idea of your product. Presentation Create a presentation on your idea using the poster making materials provided. Your presentation should be 5 minutes long, and every team member should get a chance to speak. TIP! 12
Team roles REMEMBER: You will all work on different parts of the project, but your job title shows which part of the project you will lead. You are in charge of making sure that thing gets done, but everyone on your team should have an input in all the tasks. Project Manager Software Lead Makes sure the whole team and the project is on track The creative minds behind your program - responsible for creating a flowchart for your product, like the ones from the ‘Machine Learning Now’ workshop. Research Lead Responsible for thinking about where and how to source the voluminous sets of data you will need. Research support for other team members. Risk Lead Responsible for thinking about the risk vs. utility of your product and how to manage that. How will you help people trust your product? You need to identify the risks involved and ensure that your machine learning tool will be safe and unbiased. Design Lead Marketing Lead Responsible for the physical design of the product. Responsible for developing a marketing plan and thinking about who this tool would be useful for, how and why. 13