paper-with-me

홈 › Papers

Operation-Level Early Stopping for Robustifying Differentiable NAS

2023-09-21 · NeurIPS 2023 11

Differentiable NAS (DARTS) is a simple and efficient neural architecture search method that has been extensively adopted in various machine learning tasks. % Nevertheless, DARTS still encounters several robustness issues, mainly the domination of skip connections. % The resulting architectures are full of parametric-free operations, leading to performance collapse. % Existing methods suggest that the skip connection has additional advantages in optimization compared to other parametric operations and propose to alleviate the domination of skip connections by eliminating these additional advantages. % In this paper, we analyze this issue from a simple and straightforward perspective and propose that the domination of skip connections results from parametric operations overfitting the training data while architecture parameters are trained on the validation data, leading to undesired behaviors. % Based on this observation, we propose the operation-level early stopping (OLES) method to overcome this issue and robustify DARTS without introducing any computation overhead. % Extensive experimental results can verify our hypothesis and the effectiveness of OLES.Submission Number: 11229

📄 PDF Abstract BibTeX

Code (1)

pasalab/oles 공식 구현 pytorch

Methods 이 논문이 사용한 방법론

Early Stopping Early Stopping is a regularization technique for deep neural networks that stops training when parameter updates no longer begin to yield improves on a validation set. In…
DARTS Differentiable Architecture Search (DART) is a method for efficient architecture search. The search space is made continuous so that the architecture can be optimized with…

Similar Papers 제목 키워드 기반

ROME: Robustifying Memory-Efficient NAS via Topology Disentanglement and Gradient Accumulation

2020-11-23 · ICCV 2023 1 · Xiaoxing Wang, Xiangxiang Chu, Yuda Fan, Zhexi Zhang 외

Albeit being a prevalent architecture searching approach, differentiable architecture search (DARTS) is largely hindered by its substantial memory cost since the entire supernet resides in the memory. This is where the s…

DisentanglementNeural Architecture Search

DARTS+: Improved Differentiable Architecture Search with Early Stopping

2019-09-13 · Hanwen Liang, Shifeng Zhang, Jiacheng Sun, Xingqiu He 외

Recently, there has been a growing interest in automating the process of neural architecture design, and the Differentiable Architecture Search (DARTS) method makes the process available within a few GPU days. However, t…

GPU

CHASE: Robust Visual Tracking via Cell-Level Differentiable Neural Architecture Search

2021-07-02 · Seyed Mojtaba Marvasti-Zadeh, Javad Khaghani, Li Cheng, Hossein Ghanei-Yakhdan 외

A strong visual object tracker nowadays relies on its well-crafted modules, which typically consist of manually-designed network architectures to deliver high-quality tracking results. Not surprisingly, the manual design…

Neural Architecture SearchSemantic SegmentationVisual Tracking

Making Differentiable Architecture Search less local

2021-04-21 · Erik Bodin, Federico Tomasi, Zhenwen Dai

Neural architecture search (NAS) is a recent methodology for automating the design of neural network architectures. Differentiable neural architecture search (DARTS) is a promising NAS approach that dramatically increase…

Neural Architecture Search

Learning Lipschitz Functions by GD-trained Shallow Overparameterized ReLU Neural Networks

2022-12-28 · Ilja Kuzborskij, Csaba Szepesvári

We explore the ability of overparameterized shallow ReLU neural networks to learn Lipschitz, nondifferentiable, bounded functions with additive noise when trained by Gradient Descent (GD). To avoid the problem that in th…