We needed to create an AI piece that set the scene at the entrance to the AI event, intrigued guests and gave an ice-breaker talking point.
It shows a visualisation of high dimensional data sets that learn as a user types in a word - the relationships with other words are represented in 3D visual form. The AI learns more about the relationships as more info is added to the data set.
There was a symmetry to this thinking, as people were there in person to network and learn.
The challenge was making it a learning-based AI instead of a learnt AI model, which constantly required the model to learn and develop based on the input.
We used t-SNE (t-distributed Stochastic Neighbour Embedding) to visualise high-dimensional data. As more info is added to the data set via user input, the AI learns more about the relationships and displays that data in a 3D generative form.
Discover how our innovative technology can revolutionise your environment and engage your audience. Contact us to find out more.