On June 7, our Managing Director, Elliot Fearn, took part in an exciting event at Hull’s Craven Park. The annual Humber STEM event, dedicated to inspiring and educating young minds in science, technology, engineering, and mathematics (STEM), featured Data Stream UK as a prominent contributor. Elliot led a series of workshops focused on artificial intelligence (AI), specifically exploring neural networks within the human brain and their application in AI technology.
This event brought together a diverse array of businesses, including industry leaders such as Sewell Group, Smith and Nephew, and KCOM, all united in their dedication to fostering STEM education. With an audience of 500 children, the day promised to be an informative and impactful experience.
Task 1
The workshop began with an engaging icebreaker that awoke and captivated the young audience. They were presented with a series of real images and videos alongside generative AI deepfakes, prompting the children to recognize which ones were authentic and which were fabricated (a copy of the images is below and a deepfake video of Morgan Freeman can be found in the links below if you wish to test yourself).
This thought-provoking exercise not only showcased a significant use case for AI but also highlighted the potential dangers and risks associated with deepfake technology. By challenging the children’s perceptions and encouraging critical thinking, the importance of responsible AI and the need for vigilance in a world increasingly influenced by advanced technologies was effectively conveyed.
![](https://www.datastreamuk.com/wp-content/uploads/2023/07/Obama.jpg)
![](https://www.datastreamuk.com/wp-content/uploads/2023/07/obama-2.jpg)
One real and one deepfake image of Merkel and Obama, can you spot the fake?
Task 2
Secondly, following a quick discussion on neural networks in the human brain, the workshop took an interactive turn. The children were divided into teams, fostering collaboration and hands-on learning. Their task was to construct their own neural networks using marshmallows and skewers, turning the usually abstract concept into a tangible, messy, and fun exercise. As marshmallows connected with skewers, the young participants were able to visualise how signals would traverse their custom-built neuron networks. This activity allowed the children to experience firsthand the intricate pathways of communication within neural networks. Turning complex ideas into an accessible and memorable learning experience.
Task 3
Next, the lesson took a step further, the workshop transformed the young participants into living components of a simplified neural network. Elliot orchestrated a dynamic activity by drawing out the neural network diagram on the floor. Children then played the roles of neurons and signals. Afterwards, some children became neurons, tasked with performing calculations. While others took on the role of signals, responsible for transmitting segments of information.
As the signals traversed the network, they reached the end. Where a group of children worked collaboratively to piece together the information and decipher the animal being described. This immersive exercise allowed every child to actively participate in the process of data transmission. As well as calculation, and prediction, mirroring the workings of a real-life neural network classifier. By embodying the roles of neurons and signals, the children gained a hands-on understanding of the intricate and collaborative nature of neural networks, making complex concepts accessible and relatable.
Takeaways
The workshop and entire day proved to be an enjoyable experience for all involved. Data Stream UK effectively conveyed complex AI concepts in a fun and engaging manner. It is our hope that this event has inspired students to pursue careers in AI and contribute to the field. For more information about the event, please follow the links provided below.
Links
https://www.youtube.com/watch?v=oxXpB9pSETo
https://uk.linkedin.com/in/humber-stem-47219a258
https://projects.raspberrypi.org/en/projects/alien-language/0