SEval-NAS: A Search-Agnostic Evaluation for Neural Architecture Search
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917300352843776 |
|---|---|
| author | Mih, Atah Nuh Wang, Jianzhou Nguyen, Truong Thanh Hung Cao, Hung |
| author_facet | Mih, Atah Nuh Wang, Jianzhou Nguyen, Truong Thanh Hung Cao, Hung |
| contents | Neural architecture search (NAS) automates the discovery of neural networks that meet specified criteria, yet its evaluation procedures are often hardcoded, limiting the ability to introduce new metrics. This issue is especially pronounced in hardware-aware NAS, where objectives depend on target devices such as edge hardware. To address this limitation, we propose SEval-NAS, a metric-evaluation mechanism that converts architectures to strings, embeds them as vectors, and predicts performance metrics. Using NATS-Bench and HW-NAS-Bench, we evaluated accuracy, latency, and memory. Kendall's $τ$ correlations showed stronger latency and memory predictions than accuracy, indicating the suitability of SEval-NAS as a hardware cost predictor. We further integrated SEval-NAS into FreeREA to evaluate metrics not originally included. The method successfully ranked FreeREA-generated architectures, maintained search time, and required minimal algorithmic changes. Our implementation is available at: https://github.com/Analytics-Everywhere-Lab/neural-architecture-search |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_00099 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SEval-NAS: A Search-Agnostic Evaluation for Neural Architecture Search Mih, Atah Nuh Wang, Jianzhou Nguyen, Truong Thanh Hung Cao, Hung Machine Learning Artificial Intelligence Neural and Evolutionary Computing Neural architecture search (NAS) automates the discovery of neural networks that meet specified criteria, yet its evaluation procedures are often hardcoded, limiting the ability to introduce new metrics. This issue is especially pronounced in hardware-aware NAS, where objectives depend on target devices such as edge hardware. To address this limitation, we propose SEval-NAS, a metric-evaluation mechanism that converts architectures to strings, embeds them as vectors, and predicts performance metrics. Using NATS-Bench and HW-NAS-Bench, we evaluated accuracy, latency, and memory. Kendall's $τ$ correlations showed stronger latency and memory predictions than accuracy, indicating the suitability of SEval-NAS as a hardware cost predictor. We further integrated SEval-NAS into FreeREA to evaluate metrics not originally included. The method successfully ranked FreeREA-generated architectures, maintained search time, and required minimal algorithmic changes. Our implementation is available at: https://github.com/Analytics-Everywhere-Lab/neural-architecture-search |
| title | SEval-NAS: A Search-Agnostic Evaluation for Neural Architecture Search |
| topic | Machine Learning Artificial Intelligence Neural and Evolutionary Computing |
| url | https://arxiv.org/abs/2603.00099 |