mlsanityData to 14 Oct 2026

Fair DARTS: Eliminating Unfair Advantages in Differentiable Architecture Search

arXiv 1911.12126 ↗First posted 27 Nov 2019cs.LG cs.AI cs.CVcited by 131

Not on the map. Why?

Abstract (arXiv)

Differentiable Architecture Search (DARTS) is now a widely disseminated weight-sharing neural architecture search method. However, it suffers from well-known performance collapse due to an inevitable aggregation of skip connections. In this paper, we first disclose that its root cause lies in an unfair advantage in exclusive competition. Through experiments, we show that if either of two conditions is broken, the collapse disappears. Thereby, we present a novel approach called Fair DARTS where the exclusive competition is relaxed to be collaborative. Specifically, we let each operation's architectural weight be independent of others. Yet there is still an important issue of discretization discrepancy. We then propose a zero-one loss to push architectural weights towards zero or one, which approximates an expected multi-hot solution. Our experiments are performed on two mainstream search spaces, and we derive new state-of-the-art results on CIFAR-10 and ImageNet. Our code is available on https://github.com/xiaomi-automl/fairdarts .

Key facts. Cited by 131 papers in the map (data through 14 Oct 2026). First posted on arXiv 27 Nov 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 44; 2022: 24; 2023: 19; 2024: 19; 2025: 15; 2026: 10.

131 papers.

Sort:

By area:

  1. 2023Neural Architecture Search: Insights from 1000 Paperscited by 139
  2. 2023MobileVLM : A Fast, Strong and Open Vision Language Assistant for Mobile Devicescited by 122
  3. 2021Learn to Match: Automatic Matching Network Design for Visual Trackingcited by 67
  4. 2021DiNTS: Differentiable Neural Network Topology Search for 3D Medical Image Segmentationcited by 48
  5. 2021Neural Architecture Search with Random Labelscited by 33
  6. 2021KNAS: Green Neural Architecture Searchcited by 33
  7. 2021AdvRush: Searching for Adversarially Robust Neural Architecturescited by 28
  8. 2022Riemannian Local Mechanism for SPD Neural Networkscited by 22
  9. 2021OPANAS: One-Shot Path Aggregation Network Architecture Search for Object Detectioncited by 19
  10. 2021Meta Navigator: Search for a Good Adaptation Policy for Few-shot Learningcited by 19
  11. 2021ZARTS: On Zero-order Optimization for Neural Architecture Searchcited by 18
  12. 2022Demystifying the Neural Tangent Kernel from a Practical Perspective: Can it be trusted for Neural Architecture Search without training?cited by 18
38 more
  1. 2021AutoGEL: An Automated Graph Neural Network with Explicit Link Informationcited by 17
  2. 2024Design Principle Transfer in Neural Architecture Search via Large Language Modelscited by 17
  3. 2023MSINet: Twins Contrastive Search of Multi-Scale Interaction for Object ReIDcited by 16
  4. 2021Trilevel Neural Architecture Search for Efficient Single Image Super-Resolutioncited by 15
  5. 2021Accelerating Neural Architecture Search via Proxy Datacited by 14
  6. 2024SWAP-NAS: Sample-Wise Activation Patterns for Ultra-fast NAScited by 11
  7. 2022RF-Next: Efficient Receptive Field Search for Convolutional Neural Networkscited by 10
  8. 2021Reconsidering CO2 emissions from Computer Visioncited by 9
  9. 2021RANK-NOSH: Efficient Predictor-Based Architecture Search via Non-Uniform Successive Halvingcited by 9
  10. 2022Efficient Joint-Dimensional Search with Solution Space Regularization for Real-Time Semantic Segmentationcited by 9
  11. 2023Performance-aware Approximation of Global Channel Pruning for Multitask CNNscited by 9
  12. 2021Landmark Regularization: Ranking Guided Super-Net Training in Neural Architecture Searchcited by 8
  13. 2022PRE-NAS: Predictor-assisted Evolutionary Neural Architecture Searchcited by 8
  14. 2021Neural Architecture Search for Efficient Uncalibrated Deep Photometric Stereocited by 7
  15. 2022Partial Connection Based on Channel Attention for Differentiable Neural Architecture Searchcited by 7
  16. 2023β-DARTS++: Bi-level Regularization for Proxy-robust Differentiable Architecture Searchcited by 7
  17. 2025SEKI: Self-Evolution and Knowledge Inspiration based Neural Architecture Search via Large Language Modelscited by 7
  18. 2021Learning from Mistakes -- A Framework for Neural Architecture Searchcited by 6
  19. 2023PA&DA: Jointly Sampling PAth and DAta for Consistent NAScited by 6
  20. 2021Neural Architecture Search via Combinatorial Multi-Armed Banditcited by 5
  21. 2022Learning Features with Parameter-Free Layerscited by 5
  22. 2023Efficient Automation of Neural Network Design: A Survey on Differentiable Neural Architecture Searchcited by 5
  23. 2023IS-DARTS: Stabilizing DARTS through Precise Measurement on Candidate Importancecited by 5
  24. 2024Unsupervised Graph Neural Architecture Search with Disentangled Self-supervisioncited by 5
  25. 2024CE-NAS: An End-to-End Carbon-Efficient Neural Architecture Search Frameworkcited by 5
  26. 2024Scalable Reinforcement Learning-based Neural Architecture Searchcited by 5
  27. 2021Differentiable Architecture Search for Reinforcement Learningcited by 4
  28. 2021Self-Supervised Neural Architecture Search for Imbalanced Datasetscited by 4
  29. 2021DAAS: Differentiable Architecture and Augmentation Policy Searchcited by 4
  30. 2022Differentiable Architecture Search with Random Featurescited by 4
  31. 2023Differentiable Genetic Programming for High-dimensional Symbolic Regressioncited by 4
  32. 2024Relax DARTS: Relaxing the Constraints of Differentiable Architecture Search for Eye Movement Recognitioncited by 4
  33. 2024OStr-DARTS: Differentiable Neural Architecture Search based on Operation Strengthcited by 4
  34. 2025Subnet-Aware Dynamic Supernet Training for Neural Architecture Searchcited by 4
  35. 2021Ownership Verification of DNN Architectures via Hardware Cache Side Channelscited by 3
  36. 2021Learning by Teaching, with Application to Neural Architecture Searchcited by 3
  37. 2021MANAS: Multi-Scale and Multi-Level Neural Architecture Search for Low-Dose CT Denoisingcited by 3
  38. 2021Coarse-to-Fine Searching for Efficient Generative Adversarial Networkscited by 3

50 most cited of 131 All 131 (JSON)