Optimization and training dynamics · MLD-0432
Fisher information
A matrix measuring local sensitivity of a likelihood to parameter changes in expectation.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Define a probabilistic model's local geometry.
Original caveat
Empirical and expected Fisher matrices are not always interchangeable.
Context and aliases
Level: Intermediate. Stages: Training · Theory.
Tags: ML · DL · SLM · LLM.
Aliases: None recorded.
Related concepts
Concepts linking here
Original references
- Deep Learning: online textbook ↗ · Consulted
- PyTorch optimization algorithms ↗ · Reference link
- PyTorch documentation ↗ · Reference link
- Original reference URL ↗