The visual identity for the 2021 Central Saint Martins Graduate Showcase uses artificial intelligence in the form of machine learning to represent the individual and collective growth of the graduating students within the university's community.

Collecting the word "bloom" in all languages spoken on the Central Saint Martins campus, we trained a neural network that generates new letterforms based on the initial data. Combining moving imagery, colour, and typography, we developed an adaptive visual system with this generative cycle at its core.

In collaboration with Jann Choy, Betty Lu, Jessie Zhang and Maximillian Zimmerer

Visual Identity | Artificial Intelligence | Motion Design
Media: Adobe Suite, JavaScript, Python and StyleGan
Role: Concept Development, Machine Learning, Layout and Postprocessing

The colours that the GAN is trained on are bright and celebratory, in line with the atmosphere of the degree show. As the model learns these colours, it expands on the palette and produces new variations, becoming more than the sum of its parts.

The grid system, which is based on the 24 courses exhibiting as part of the Degree Show, is the foundation of the identity system.

The visual language is centered around a square moving image of GAN-generated letterforms. Striking a balance between order and chaos, we establish a system that is easily transformable to any platform.

Design Development

Upon reflection of our time at Central Saint Martins, we concluded that the Graduate Showcase represented how we have grown over the years, both as individuals and as an entire collective. Through the process of interaction and collaboration with our peers and tutors, we grow, together.

The learning process that a GAN goes through is analagous to the way in which we learn as students. Throughout the journeys that we undertake at Central Saint Martins, we absorb knowledge from our tutors and peers as we carve our our path, forming our own unique practices.

We collected over 4000 typefaces from the internet as the starting point for our identity, using these to train a GAN, which over time learnt the shapes of these glyps and created its own.

Working Process

Next Project

Water to Water

© Thomas Bugg 2023 → ∞