SECAI Core · phase 2 of 15 · noun
Pruning
Removing less important elements of the model that have negligible impact on performance
The Explain card
- Plain English
- Pruning removes parts of a model, such as weights or neurons, that contribute almost nothing to its output, making it smaller and faster with little loss in performance.
- Example
- An engineer prunes a vision model by half so it fits on an edge camera, then re-tests accuracy to confirm detection still works.
- Why it matters
- Pruning changes behaviour in subtle ways. It can strip safety-relevant capacity, and some attacks design a backdoor that only activates after a model is pruned or compressed.
- Hook
- Trim the dead branches, but check what was living in them.
Word knowledge
How the word is built, where it came from, and what it sits beside in memory.
In a sentence
Magnitude pruning is run on the trained convolutional network so the detector fits the doorbell camera's on-chip SRAM.
How it is built
- prune root to cut back, trim
- -ing suffix the action or its result
Where it came from
- Origin
- Middle English prouynen, from Old French proignier, to cut back a vine, of uncertain further origin, perhaps Vulgar Latin *provineare, to dress a vine, from Latin pro plus vinea, vineyard
- Entered the language
- late 14th century
- What changed
- English gardeners used the verb for cutting back vines, then wrote prune, with pruning as the gerund. Birds kept a related sense as preen. Decision-tree papers of the 1980s applied the verb to learned trees, and neural-network papers of the 1990s applied it to trained weights.
How it is spelled
- Pattern
- A silent e drops before a vowel suffix: prune becomes pruning, never pruneine
- Pattern
- u-e spells the long /uː/ in prune, as in June and flute
- Breaks the pattern
- The dried-plum prune is a homograph from Latin prunum, not this family
- Breaks the pattern
- Middle English spelled the verb prouynen and proinen before the modern silent-e form settled
Spelled like
- tuning
- coding
- tracing
- preening
Broken into chunks
-
prune
- pruned
- pruner
- preen
-
ing
- training
- tuning
- learning
What it sits beside
Same subject
- quantization
- knowledge distillation
- sparsity
- fine tuning
Same shape
- tuning
- preening
- proving
- grouping
Model compression
- quantization
- knowledge distillation
- sparsity
Gerund process nouns
- training
- fine tuning
- embedding
Silent-e verbs that drop e before -ing
- tuning
- coding
- tracing
Where it sits in the deck
Phase 2: How Models Are Built: Architectures and Learning Mechanics
With the paradigms named, zoom in on the architectural building blocks that turn data into trained artifacts.
