Weight-Entanglement Meets Gradient-Based Neural Architecture Search

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sukthanker, Rhea Sanjay, Krishnakumar, Arjun, Safari, Mahmoud, Hutter, Frank
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915607562158080
author Sukthanker, Rhea Sanjay
Krishnakumar, Arjun
Safari, Mahmoud
Hutter, Frank
author_facet Sukthanker, Rhea Sanjay
Krishnakumar, Arjun
Safari, Mahmoud
Hutter, Frank
contents Weight sharing is a fundamental concept in neural architecture search (NAS), enabling gradient-based methods to explore cell-based architectural spaces significantly faster than traditional black-box approaches. In parallel, weight-entanglement has emerged as a technique for more intricate parameter sharing amongst macro-architectural spaces. Since weight-entanglement is not directly compatible with gradient-based NAS methods, these two paradigms have largely developed independently in parallel sub-communities. This paper aims to bridge the gap between these sub-communities by proposing a novel scheme to adapt gradient-based methods for weight-entangled spaces. This enables us to conduct an in-depth comparative assessment and analysis of the performance of gradient-based NAS in weight-entangled search spaces. Our findings reveal that this integration of weight-entanglement and gradient-based NAS brings forth the various benefits of gradient-based methods, while preserving the memory efficiency of weight-entangled spaces. The code for our work is openly accessible https://github.com/automl/TangleNAS.
format Preprint
id arxiv_https___arxiv_org_abs_2312_10440
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Weight-Entanglement Meets Gradient-Based Neural Architecture Search
Sukthanker, Rhea Sanjay
Krishnakumar, Arjun
Safari, Mahmoud
Hutter, Frank
Machine Learning
Artificial Intelligence
Weight sharing is a fundamental concept in neural architecture search (NAS), enabling gradient-based methods to explore cell-based architectural spaces significantly faster than traditional black-box approaches. In parallel, weight-entanglement has emerged as a technique for more intricate parameter sharing amongst macro-architectural spaces. Since weight-entanglement is not directly compatible with gradient-based NAS methods, these two paradigms have largely developed independently in parallel sub-communities. This paper aims to bridge the gap between these sub-communities by proposing a novel scheme to adapt gradient-based methods for weight-entangled spaces. This enables us to conduct an in-depth comparative assessment and analysis of the performance of gradient-based NAS in weight-entangled search spaces. Our findings reveal that this integration of weight-entanglement and gradient-based NAS brings forth the various benefits of gradient-based methods, while preserving the memory efficiency of weight-entangled spaces. The code for our work is openly accessible https://github.com/automl/TangleNAS.
title Weight-Entanglement Meets Gradient-Based Neural Architecture Search
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2312.10440