# Training Data
> **Machine-Evolution Nexus:** [[articles/Digital Darwinism and the Invisible World of Machine Evolution|Digital Darwinism and the Invisible World of Machine Evolution]] places this node within the history and governance of substrate-independent evolution.
**Domain:** Machine Learning / Provenance
**Doc Type:** Canonical Wiki Node
**Maturity:** Developed
## Definition
Training data are examples or observations used to fit a machine-learning model’s parameters or behavior.
## Nexus Context
For protein prediction, biological databases contain sequences and structures shaped by evolutionary history. Evolution is therefore part of the data-generating process, not merely an analogy.
## Evidence Boundary
Data are selected, measured and curated. Provenance, coverage, leakage, consent and historical bias constrain what a model’s performance means.
## Relationships
[[wiki/Machine Learning|Machine Learning]], [[wiki/Provenance|Provenance]], [[wiki/AlphaFold|AlphaFold]], [[wiki/Public Data Infrastructure|Public Data Infrastructure]]
- **Article:** [[articles/war with empire/Authorship of the West|Authorship of the West]] argues that every dataset carries civilization's disputes, because the disputes are part of the dataset.
## Simple Reminders, Quotations, and Thoughts
> “Artificial intelligence will inherit more than our information; it will inherit the arguments we leave unresolved.”
> **— Bryant McGill**, *Authorship of the West, September 2026*
[[reminders/Machine Succession/AI Will Inherit the Arguments We Leave Unresolved by Bryant McGill|AI Will Inherit the Arguments We Leave Unresolved by Bryant McGill]]
> “The machines of the future will know almost everything about our civilization, but knowledge alone will not tell them what was worth preserving.”
> **— Bryant McGill**, *Authorship of the West, September 2026*
[[reminders/Preservation/Machines Will Need More Than Knowledge to Know What Was Worth Preserving by Bryant McGill|Machines Will Need More Than Knowledge to Know What Was Worth Preserving by Bryant McGill]]
## Sources / Provenance
- Machine-learning methodology; AlphaFold methods.