Index.training
| created and maintained by Chaski | 2024 |
|---|
Description
index.training explores the index as a creative infrastructure for mapping knowledge and associative relationships within information. Using Natural Language Processing (NLP), it visualizes a dataset of index entries drawn from three texts selected by the artist from the Cybernetics Library’s Uncomputable collection:
- Updating to Remain the Same: Habitual New Media — Wendy Hui Kyong Chun
- Code and Clay, Data and Dirt: Five Thousand Years of Urban Media — Shannon Christine Mattern
- The Moment of Complexity: Emerging Network Culture — Mark C. Taylor
Visitors interact with a 3D t-SNE visualization where each point represents an index entry, encountering a multiply-determined, interactive topology of ideas. Each entry’s position is determined by its semantic proximity to others in the dataset, as calculated by an approximate nearest neighbors algorithm. Rather than aiming to produce correct or logical groupings of items—which it rarely does—the project is more interested in allowing the spatial layout of text on the page to be shaped by another way of mapping knowledge.