A Meta-Learning Approach for Custom Model Training
Explore this paper's citation graph
Summary
This paper proposes a joint training approach that combines both transfer-learning and meta-learning, and obtains improved generalization performance on unseen target tasks in both few- and many-class and few-and-many-shot scenarios.
- Type
- preprint
- Published
- 2018-09-21
- Cited by
- 7
- References
- 10
- Access
- Open access
- OpenAlex
- https://openalex.org/W2890884924
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:52811286
Keywords
Meta learning (computer science), Computer science, Transfer of learning, Artificial intelligence, Generalization
References
- A Survey on Transfer Learning
- Prototypical Networks for Few-shot Learning
- Optimization as a Model for Few-Shot Learning
- On First-Order Meta-Learning Algorithms
- Gradient Agreement as an Optimization Objective for Meta-Learning
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Matching Networks for One Shot Learning
- How transferable are features in deep neural networks?
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
Cited by
- Efficient Training of Deep Convolutional Neural Networks by Augmentation in Embedding Space
- A survey on data‐efficient algorithms in big data era
- Face presentation attack detection. A comprehensive evaluation of the generalisation problem
- Context vector-based visual mapless navigation in indoor using hierarchical semantic information and meta-learning
- Task-based Meta Focal Loss for Multilingual Low-resource Speech Recognition
- Self-learning UAV Motion Planning Based on Meta Reinforcement Learning
- A cross-domain few-shot remaining useful life estimation framework based on model-agnostic meta-learning with task embeddings
Related papers
- Few-shot Learning with Meta Metric Learners
- Variable-Shot Adaptation for Online Meta-Learning
- A New Meta-Baseline for Few-Shot Learning
- A Hybrid Approach with Optimization and Metric-based Meta-Learner for Few-Shot Learning
- Meta-Meta Classification for One-Shot Learning
- A hybrid approach with optimization-based and metric-based meta-learner for few-shot learning