Saved in:
Bibliographic Details
Main Author: He, Yannis Y.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.02007
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913634992521216
author He, Yannis Y.
author_facet He, Yannis Y.
contents In the realm of neural architecture design, achieving high performance is largely reliant on the manual expertise of researchers. Despite the emergence of Neural Architecture Search (NAS) as a promising technique for automating this process, current NAS methods still require human input to expand the search space and cannot generate new architectures. This paper explores the potential of Transformers in comprehending neural architectures and their performance, with the objective of establishing the foundation for utilizing Transformers to generate novel networks. We propose the Token-based Architecture Transformer (TART), which predicts neural network performance without the need to train candidate networks. TART attains state-of-the-art performance on the DeepNets-1M dataset for performance prediction tasks without edge information, indicating the potential of Transformers to aid in discovering novel and high-performing neural architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2501_02007
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TART: Token-based Architecture Transformer for Neural Network Performance Prediction
He, Yannis Y.
Machine Learning
Artificial Intelligence
In the realm of neural architecture design, achieving high performance is largely reliant on the manual expertise of researchers. Despite the emergence of Neural Architecture Search (NAS) as a promising technique for automating this process, current NAS methods still require human input to expand the search space and cannot generate new architectures. This paper explores the potential of Transformers in comprehending neural architectures and their performance, with the objective of establishing the foundation for utilizing Transformers to generate novel networks. We propose the Token-based Architecture Transformer (TART), which predicts neural network performance without the need to train candidate networks. TART attains state-of-the-art performance on the DeepNets-1M dataset for performance prediction tasks without edge information, indicating the potential of Transformers to aid in discovering novel and high-performing neural architectures.
title TART: Token-based Architecture Transformer for Neural Network Performance Prediction
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2501.02007