About
This project is an algorithm for walking. A set of instructions to be carried out as an action, transposing the logic of the computer algorithm of ‘Gradient Descent’ (for moving in an abstract mathematical space), into the realm of physical reality. Using the affordances of language and terminology to cross over different fields, and the capacity to be taken very literally, the ‘Gradient Descend’ becomes a proposition for adopting the logic of a computer algorithm to walk down a hill. The different individual actions are collected together into a virtual atlas of descents
About the algorithm
‘Gradient descent’ is a method for unconstrained mathematical optimization. It is a first-order iterative algorithm for finding a local minimum of a differentiable multivariate function. It is used for data optimisation, prediction and training of machine learning models. It takes small steps on a mathematical function to arrive to the lowest point, in the direction with the steepest slope.
Algorithms ~ walking?
By bringing these two clashing logics of algorithmic and bodily movement, together, ( clashing because it feels almost dangerous to walk down a mountain in the steepest direction ), some questions open up:
- How do maps and algorithms embedded in mobile devices, direct the way we move in space.
- How can the process of mapping as a method register individual motion, and choice, instead of prescribing it?
- How is optimisation and prediction used to find the ‘best routes’ to take?
- What new paths can emerge our of these directions, if we see paths as algorithms inscribed on the ground?
____________________________
This project is part of my ongoing research into the epistemologies and spatial logics of algorithmic and bodily movements in space, and the way they are registered or prescribed through the process of mapping.
____________________________