Reference library
Net2Net: Accelerating Learning via Knowledge Transfer
Paper summary
Net2Net expands neural networks through function-preserving widening and deepening transformations. Transferring an existing solution to a larger model accelerates subsequent training compared with starting from scratch.
Relevance to the project
A candidate comparison for changing mechanism-network capacity without immediate functional disruption. Builder evolution and mutable graph growth still require distinct rules.
Preserving a feedforward function does not guarantee identical recurrent adaptive trajectories.
Source record
- Reference ID
- net2net
- Archived version
- 1511.05641v4
- Summary provenance
- reference-library/papers/net2net/metadata.json · summary
- Review status
- initial orientation summary; not a critical full-paper review
- Reading scope
- Selected abstract/opening and available closing discussion/conclusion passages in the preserved text extract; broader project notes provide context. No reproduction of experiments, exhaustive methods review, or equation/figure audit.