# Disturbance-immune Weight Sharing for Neural Architecture Search

arXiv: https://arxiv.org/abs/2003.13089 · HTML: https://mlsanity.com/paper/2003.13089/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Mar 2020.

## Built on by (most cited first, 9 of 9)

- [Efficient Test-Time Model Adaptation without Forgetting](https://mlsanity.com/paper/2204.02610/) (2022)
- [EPMF: Efficient Perception-aware Multi-sensor Fusion for 3D Semantic Segmentation](https://mlsanity.com/paper/2106.15277/) (2021)
- [Towards Accurate Text-based Image Captioning with Content Diversity Exploration](https://mlsanity.com/paper/2105.03236/) (2021)
- [Uncertainty-Calibrated Test-Time Model Adaptation without Forgetting](https://mlsanity.com/paper/2403.11491/) (2024)
- [CLOSE: Curriculum Learning On the Sharing Extent Towards Better One-shot NAS](https://mlsanity.com/paper/2207.07868/) (2022)
- [Boosting Semi-Supervised Learning with Contrastive Complementary Labeling](https://mlsanity.com/paper/2212.06643/) (2022)
- [FIESTA: Fisher Information-based Efficient Selective Test-time Adaptation](https://mlsanity.com/paper/2503.23257/) (2025)
- [Neural Architecture Retrieval](https://mlsanity.com/paper/2307.07919/) (2023)
- [Learning to Rank Ace Neural Architectures via Normalized Discounted Cumulative Gain](https://mlsanity.com/paper/2108.03001/) (2021)

## Abstract (arXiv)

Neural architecture search (NAS) has gained increasing attention in the community of architecture design. One of the key factors behind the success lies in the training efficiency created by the weight sharing (WS) technique. However, WS-based NAS methods often suffer from a performance disturbance (PD) issue. That is, the training of subsequent architectures inevitably disturbs the performance of previously trained architectures due to the partially shared weights. This leads to inaccurate performance estimation for the previous architectures, which makes it hard to learn a good search strategy. To alleviate the performance disturbance issue, we propose a new disturbance-immune update strategy for model updating. Specifically, to preserve the knowledge learned by previous architectures, we constrain the training of subsequent architectures in an orthogonal space via orthogonal gradient descent. Equipped with this strategy, we propose a novel disturbance-immune training scheme for NAS. We theoretically analyze the effectiveness of our strategy in alleviating the PD risk. Extensive experiments on CIFAR-10 and ImageNet verify the superiority of our method.