.png){kind=icon}
.png){kind=icon}
Return
A Continuous Encoding-Based Representation for Efficient Multi-Fidelity Multi-Objective Neural Architecture Search
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2025.113932.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A continuous encoding method is proposed to reduce the number of variables in the representation of generalized U-Net-based neural architecture search (NAS). • An adaptive Co-Kriging-assisted multi-fidelity multi-objective NAS algorithm is proposed to decrease the computational costs. • The proposed NAS algorithm identifies superior U-Net architectures that leverage information flow from previous cells.
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W