본문 바로가기
KERD
논문KCI 등재

이미지 분석을 위한 퓨샷 학습의 최신 연구동향

Recent advances in few-shot learning for image domain: a survey

석호식
한국전기전자학회 (Institute of Korean Electrical and Electronics Engineers)|2023.12|전기전자학회논문지 (Journal of IKEEE)|제27권 제4호|pp. 537-547|DOI 10.7471/ikeee.2023.27.4.537

2

조회

0

다운로드

0

인용

0

참고문헌

키워드

machine learningzero-shot learningfew-shot learningone-shot learningimage classification.
주제분류전자/정보통신공학

초록

퓨삿학습(few-shot learning)은 사전에 확보한 관련 지식과 소규모의 학습데이터를 이용하여 학습데이터의 부족으로 인한 어려움을 해결할 수 있는 가능성을 제시해주어 최근 많은 주목을 받고 있다. 본 논문에서는 퓨삿학습의 개념과 주요 접근방법을 빠르게파악할 수 있도록 데이터 증강, 임베딩과 측도학습, 메타학습의 세 관점에서 최신연구동향을 설명한다. 또한 퓨샷학습을 적용하려는연구자들에게 도움을 제공할 수 있도록 주요 벤치마크 데이터셋에 대하여 간략하게 소개하였다. 퓨삿학습은 이미지 분석과 자연어처리 등 다양한 분야에서 활용되고 있으나, 본 논문은 이미지 처리를 위한 퓨삿학습의 접근법에 집중하였다

Abstract

In many domains, lack of data inhibits adoption of advanced machine learning models. Recently, Few-ShotLearning (FSL) has been actively studied to tackle this problem. Utilizing prior knowledge obtained throughobservations on related domains, FSL achieved significant performance with only a few samples. In this paper, wepresent a survey on FSL in terms of data augmentation, embedding and metric learning, and meta-learning. Inaddition to interesting researches, we also introduce major benchmark datasets. FSL is widely adopted in variousdomains, but we focus on image analysis in this paper.

참고문헌 (55)

  1. Yaqing Wang, "Generalizing from a Few Examples", ACM Computing Surveys, 53(3), pp.1~34 (2020), Association for Computing Machinery (ACM)[DOI]
  2. S. Laenen, "On Episodes, Prototypical Networks, and Few-Shot Learning", Proc. of 35th Conferenceon Neural Information Processing Systems (NeurIPS2021) (2021)
  3. A. Rajeswaran, "Meta-Learning with Implicit Gradients", Proc. of 33rd Conference on Neural Information Processing Systems (NeurIPS2019) (2019)
  4. Connor Shorten, "A survey on Image Data Augmentation for Deep Learning", Journal of Big Data, 6(1) (2019), Springer Science and Business Media LLC[DOI]
  5. Rui Zhang, "Multi‐task few‐shot learning with composed data augmentation for image classification", IET Computer Vision, 17(2), pp.211~221 (2022), Institution of Engineering and Technology (IET)[DOI]
  6. V. Verma, "Manifold Mixup: Better Representations by Interpolating Hidden States", Proc. of the Thirty-sixth International Conference on Machine Learning (ICML2019) (2019)
  7. H. Zhang, "mixup: Beyond Empirical Risk Minimization", Proc. of International Conference Learning Representation (ICLR 2018) (2018)
  8. L. Zhao, "Maximum-Entropy Adversarial Data Augmentation for Improved Generalization and Robustness", Proc. of the 34th Conference on Neural Information Processing Systems (Neur-IPS2020) (2020)
  9. C. Gong, "MaxUp : Lightweight Adversarial Training with Data Augmentation Improves Neural Network Training", Proc. 2021 IEEE Conference on Computer Vision and Pattern Recognition (CVPR2021), pp.2474~2483 (2021)
  10. P. T. Jackson, "Style Augmentation:Data Augmentation via Style Randomization", CVPR Workshop 2019, pp.83~92 (2019)
  11. Yuchen Jiang, "Remote Sensing Images Data Augmentation Based on Style Transfer under the Condition of Few Samples", Journal of Physics: Conference Series, 1653(1), pp.012039~ (2020), IOP Publishing[DOI]
  12. Jennifer Hemmerich, "COVER: conformational oversampling as data augmentation for molecules", Journal of Cheminformatics, 12(1) (2020), Springer Science and Business Media LLC[DOI]
  13. Liangsheng Wu, "Data augmentation based on multiple oversampling fusion for medical image segmentation", PLOS ONE, 17(10), pp.e0274522~ (2022), Public Library of Science (PLoS)[DOI]
  14. A. Moreo, "Distributional Random Oversampling for Imbalanced Text Classification", Proc. of the 39th International ACM SIGIR conference on Research and Development in Information (SIGIR2016), pp.805~808 (2016)
  15. A. Anand, "Phishing URL Detection with Oversampling based on Text Generative Adversarial Networks", Proc. IEEE International Conference on Big Data (2018)
  16. N. V. Chawla, "SMOTE: Synthetic Minority Over-sampling Technique", Journal of Artificial Intelligence Research, 16, pp.321~357 (2002), AI Access Foundation[DOI]
  17. Mateusz Ochal, "Few-Shot Learning With Class Imbalance", IEEE Transactions on Artificial Intelligence, 4(5), pp.1348~1358 (2023), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
  18. D. Wertheimer, "Few-Shot Learning with Localization in Realistic Settings", Proc. 2019 IEEE, 2019, 6558~6567
  19. Brian Kulis, "Metric Learning: A Survey", Foundations and Trends® in Machine Learning, 5(4), pp.287~364 (2013), Now Publishers[DOI]
  20. A Survey on Metric Learning for Feature Vectors and Structured Data, "A. Bellet", ~ (2013)
  21. D. Kedem, "Non-Linear Metric Learning", Proc. of Advances in Neural Information Processing Systems 25 (NIPS2012) (2012)
  22. M. Kaya, "Deep Metric Learning: A Survey", Symmetry, 11(9), pp.1066~ (2019), MDPI AG[DOI]
  23. O. Vinyals, "Matching Networks for One Shot Learning", Proc. of the 30th Conference on Neural Information Processing Systems (NIPS2016) (2016)
  24. F. Sung, "Learning to Compare: Relation Network for Few-Shot Learning", Proc. of 2018 IEEE, 2018, 1199~1208
  25. H.-J. Ye, "Few-Shot Learning via Embedding Adaptation with Set-to-Set Functions", Proc. of IEEE, 2020, 8808~8817
  26. C. Liu, "Learning a Few-shot Embedding Model with Contrastive Learning", Proc. of The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21), pp.8635~8643 (2021)
  27. Y. Gao, "Contrastive Prototype Learning with Augmented Embeddings for Few-Shot Learning", Proc. of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence (UAI 2021), pp.140~150 (2021)
  28. P. Rodriguez, "Embedding Propagation:Smoother Manifold for Few-Shot Classication", Proc. of European Conference on Computer Vision (ECCV2020), pp.121~138 (2020)
  29. Shiming Xiang, "Learning a Mahalanobis distance metric for data clustering and classification", Pattern Recognition, 41(12), pp.3600~3612 (2008), Elsevier BV[DOI]
  30. H.-J. Ye, "Learning Mahalanobis Distance Metric: Considering Instance Disturbance Helps", Proc. of the 2017 International Joint Conference on Artificial Intelligence (IJCAI-17), pp.3315~3321 (2017)
  31. J. Goldberger, "Neighbourhood Components Analysis", Proc. of Advances in Neural Information Processing Systems 17 (NIPS 2004) (2004)
  32. J. Snell, "Prototypical Networks for Few-Shot Learning", Proc. of the 31st Conference on Neural Information Processing Systems (NIPS2017) (2017)
  33. P. Bateni, "Improved Few-Shot Visual Classification", Proc. of 2020 IEEE, 2020, 14481~14490
  34. W. Li, "Distribution Consistency based Covariance Metric Networks for Few-Shot Learning", Proc. of the Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19), pp.8642~8649 (2019)
  35. L. Qiao, "Transductive Episodic-wise Adaptive Metric for Few-Shot Learning", Proc. of 2019 IEEE, 2019, 3603~3612
  36. Timothy M Hospedales, "Meta-Learning in Neural Networks: A Survey", IEEE Transactions on Pattern Analysis and Machine Intelligence, 44, pp.5149~5169 (2021), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
  37. C. Finn, "Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks", Proc. of the 34th International Conference on Machine Learning (ICML2017), pp.1126~1135 (2017)
  38. A. A. Rusu, "Meta-Learning with Latent Embedding Optimization", Proc. of International Conference Learning Representation (ICLR 2019) (2019)
  39. H. Yao, "Hierarchically Structured Meta-Learning", Proc. of the 36th International Conference on Machine Learning (ICML2019), pp.7045~7054 (2019)
  40. C. Fifty, "Efficiently Identifying Task Groupings for Multi-Task Learning", Proc. of the Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS2021) (2021)
  41. A. Zhou, "Meta-Learning Symmetries by Reparameterization", Proc. of International Conference Learning Representation (ICLR 2021) (2021)
  42. S.-O. Kaba, "Equivariance with Learned Canonicalization Functions", Proc. of the 40th International Conference on Machine Learning (ICML2023) (2023)
  43. S. Basu, "Equivariant Few-Shot Learning from Pretrained Models", arXiv:2305.09900, ~ (2023)
  44. E. Triantafillou, "Learning a Universal Template for Few-shot Dataset Generalization", Proc. of the 38th International Conference on Machine Learning (ICML2021), pp.10424~10433 (2021)
  45. Vincent Dumoulin, "Feature-wise transformations", Distill, 3(7) (2018), Distill Working Group[DOI]
  46. M. Andrychowicz, "Learning to Learn by Gradient Descent by Gradien Descent", Proc. of the 30th Conference on Neural Information Processing Systems (NIPS2016) (2016)
  47. K. Li, "Learning to Optimize", Proc. of International Conference Learning Representation (ICLR 2017) (2017)
  48. S. Ravi, "Optimization as a Model for Few-Shot Learning", Proc. of International Conference Learning Representation (ICLR 2017) (2017)
  49. T. Chen, "Learning to Optimize: a Primer and a Benchmark", J Mach Learn Res, 23(1), pp.8562~8620 (2023)
  50. Z. Yue, "Interventional Few-Shot Learning", Proc. of the Thirty-fourth Conference on Neural Information Processing Systems (Neur IPS2020) (2020)
  51. C. Finn, "Online Meta-Learning", Proc. of the 36th International Conference on Machine Learning (ICML2019), pp.1920~1930 (2019)
  52. The Caltech-UCSD Birds-200-2011 Dataset, "C. Wah", California Institute of Technology, ~, pp.Technical Report CNS-TR-2011-001 (2011)
  53. B. M. Lake, "Human-level Concept Learning through Probabilistic Program Induction", Science, 350(6266), pp.1332~1338 (2015)
  54. A. Shaban, "One-Shot Learning for Semantic Segmentation", Proc. of British Machine Vision Conference (BMVC2017) (2017)
  55. E. Triantafillou, "Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examples", Proc. of International Conference Learning Representation (ICLR 2020) (2020)

관련 논문AI추천

Support set의 중앙값 prototype을 활용한 few-shot 학습백으뜸(사)한국스마트미디어학회 · 스마트미디어저널2023 · 제12권 제1호스케치 기반 이미지 검색의 최신 연구 동향석호식, 오세홍한국전기전자학회 · 전기전자학회논문지2024 · 제28권 제2호영상 내 물체 검출 및 분류를 위한 소규모 데이터 확장 기법김성신, 김은경, 김진용한국로봇학회 · 로봇학회 논문지2020 · 제15권 제2호인공지능 데이터 학습 기법의 일반화 성능 분석에 관한 연구조선문한국인공지능융합기술학회 · 인공지능융합기술연구2025 · 제5권 제2호퓨샷 학습과 삼중항 손실을 이용한 지상 무기체계 분류 모델의 확장성 연구배기민, 이장형화랑대연구소 · 한국군사학논집2023 · 제79권 제3호위성영상 이미지를 활용한 연구 동향 및 데이터셋 리뷰강주영, 김세형, 채정우(사)한국스마트미디어학회 · 스마트미디어저널2022 · 제11권 제1호

이 간행물의 다른 논문 (42건)

차량용 LED 매트릭스 헤드램프 제어를 위한LED 제어 프로토콜 설계 및 제어기 구현김원채, 양성현, 이창민, 이성수pp. 368-378실시간 차량 밀도에 대응하는 심층강화학습기반 C-V2X 분산혼잡제어조한신, 양우열, 전병철pp. 379-385실리콘 나노시트 피드백 전계효과 트랜지스터의준비휘발성 메모리 특성 연구허효주, 조경아, 김상식, 류승호pp. 386-390Regulation control of a dc motor by ϵ-PID controller based on the frequency response analysisAlhassan Dodo Adamou Soudeize, 박지선, 최호림pp. 391-3983D NAND Flash Memory에 Ferroelectric Material을사용한 Current Path 개선이재우, 이지환, 강명곤pp. 399-404
원문 보기