Neural Architecture Search
Definition
Architecture search automatically searches for suitable model architectures.
The automated search for model architectures in a defined search space based on an evaluation and search procedure.
Simply explained
A procedure tries out possible model architectures and compares them according to defined criteria.
Why it matters in practice
Helps realistically assess model quality, limits, and technical decisions.
Example from business
For an edge device, networks are searched based on accuracy, latency, and memory requirements.
Limits and pitfalls
- Ignoring search costs, data leakage, and hardware dependency or presenting the found architecture as universally optimal.
Origin & history
NAS became widespread from the mid-2010s through reinforcement learning, evolution, and differentiable search.
Use in mid-sized companies
- Possible application: For an edge device, networks are searched based on accuracy, latency, and memory requirements.
Frequently asked questions
What does neural architecture search mean?
The automated search for model architectures in a defined search space based on an evaluation and search procedure.
Which mistakes should be avoided?
Ignoring search costs, data leakage, and hardware dependency or presenting the found architecture as universally optimal.
Sources
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: