Support set의 중앙값 prototype을 활용한 few-shot 학습
Few-shot learning using the median prototype of the support set
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
메타 학습(meta learning)이란 즉각적으로 아는 것과 모르는 것을 구별하는 메타 인지로 적은 양의 데이터로 스스로 학습하고, 학습한 정보와 알고리즘으로 새로운 문제에 적응하며 해결하는 학습 방식이다. 그 중, few-shot 학습 방법은 메타 학습 방법의 한 종류로 매우 적은 학습 데이터 (support set)으로도 질의 데이터(query set)를 올바르게 예측하도록 하는 학습 방법이다. 본 연구에서는 각 클래스의 mean-point vector로 생성한 프로토타입의 한계점인 높은 밀도값을 낮추면서 이상치(outlier)값을 극복하는 방법을 제안한다. 제안한 방법은 기존의 방법을 해결하기 위해, 딥러닝 모델에서 feature를 추출하고, 획득한 feature사이의 요소별로 중앙값 계산하여 프로토타입을 생성하는 방법을 사용한다. 그 후, 앞서 생성한 중앙값 프로토타입을 기반으로 few-shot 학습 방법에 사용한다. 제안한 방법의 정량적인 평가를 위해 필체 인식 데이터셋을 사용하여 기존의 방법과 비교하였다. 실험 결과를 통해 기존의 방법보다 향상된 성능을 내는 것을 확인할 수 있었다.
Abstract
Meta-learning is metacognition that instantly distinguishes between knowing and unknown. It is a learning method that adapts and solves new problems by self-learning with a small amount of data.A few-shot learning method is a type of meta-learning method that accurately predicts query data even with a very small support set. In this study, we propose a method to solve the limitations of the prototype created with the mean-point vector of each class. For this purpose, we use the few-shot learning method that created the prototype used in the few-shot learning method as the median prototype. For quantitative evaluation, a handwriting recognition dataset and mini-Imagenet dataset were used and compared with the existing method. Through the experimental results, it was confirmed that the performance was improved compared to the existing method.
참고문헌 (30)
- F. Li, "One-shot learning of object categories", IEEE transactions on pattern analysis and machine intelligence, 28(4), pp.594~611 (2006)
- J. Schmidhuber, "Evolutionary principles in self-referential learning, or on learning how to learn: the meta-meta-... hook", pp.1987
- W. Lilian, "Meta-Learning: Learning to Learn Fast" (2019), OpenAI
- S. Flood, "Learning to compare: Relation network for few-shot learning", IEEE Conference on Computer Vision and Pattern Recognition, pp.1199~1208 (2018)
- S. Qianru, "Meta-transfer learning for few-shot learning", IEEE Conference on Computer Vision and Pattern Recognition, pp.403~412 (2019)
- S. Gidaris, "Dynamic few-shot visual learning without forgetting", IEEE Conference on Computer Vision and Pattern Recognition, pp.4367~4375 (2018)
- X. Wang, "Tafe-net: Task-aware feature embeddings for low shot learning", IEEE Conference on Computer Vision and Pattern Recognition, pp.1831~1840 (2019)
- C. Finn, "Modelagnostic meta-learning for fast adaptation of deep networks", International Conference on Machine Learning, 70, pp.1126~1135 (2017)
- 박대경, "Few-Shot Learning을 사용한 호스트 기반 침입 탐지 모델", 정보처리학회 논문지, 10(7), pp.271~278 (2021), 한국정보처리학회
- S. Ravi, "Optimization as a model for few-shot learning", Learning Representations(ICLR) (2017)
- T. Munkhdalai, "Rapid adaptation with conditionally shifted neurons" (2017)
- E. Grant, "Recasting gradient-based meta-learning as hierarchical bayes" (2018)
- A. Andrei, "Meta-learning with latent embedding optimization", Learning Representations (2019)
- F. Chelse, "Model agnostic meta-learning for fast adaptation of deep networks", International Conference on Machine Learning, 70, pp.1126~1135 (2017)
- O. Vinyals, "Matching networks for one shot learning", Advances in neural information processing systems, pp.3630~3638 (2016)
- J. Snell, "Prototypical networks for few-shot learning", ~ (2017)
- C. Davide, "Artificial Neural Networks, Methods in Molecular Biology, Vol. 2190", pp.73~94 (2020), Springer Protocols, Humana Press
- B. Jane, "Signature verification using a "Siamese" time delay neural network", Advances in Neural Information Processing Systems, 6, pp.737~744 (1994)
- G. Chechik, "Large Scale Online Learning of Image Similarity Through Ranking", Journal of Machine Learning Research, 11, pp.1109~1135 (2010)
- S. Hsiao, "Malware image classification using one-shot learning with Siamese networks", Procedia Computer Science, 159, pp.1863~1871 (2019)
- Y. Taigman, "Deepface: Closing the gap to human-level performance in face verification", IEEE Conference on Computer Vision and Pattern Recognition (2014)
- S. Ravi, "Optimization as a model for few-shot learning", Learning Representations(ICLR) (2017)
- W. Jeremy, "Spring Research Presentation: A Theoretical Foundation for Inductive Transfer", pp.2007
- H. Kaiming, "Rethinking imagenet pre-training", IEEE, 2019, ~
- H. Kaiming, "Deep residual learning for image recognition", IEEE conference on computer vision and pattern recognition (2016)
- V. Oriol, "Matching networks for one shot learning", Advances in neural information processing systems, 29 (2016)
- B. Lake, "Human-level concept learning through probabilistic program induction", Science, 350(6266), pp.1332~1338 (2015)
- B. Lake, "Human-level concept learning through probabilistic program induction", Science, 350(6266), pp.1332~1338 (2015)
- I. J. Good, "Some Terminology and Notation in Information Theory", Proceedings of the IEEE-Part C : Monographs, 103(3), pp.200~204 (1956)
- D. Kingma, "Adam : A method for stochastic optimization" (2014)