Uncertainty in Multitask Transfer Learning
arXiv:1806.07528
Abstract
Using variational Bayes neural networks, we develop an algorithm capable of accumulating knowledge into a prior from multiple different tasks. The result is a rich and meaningful prior capable of few-shot learning on new tasks. The posterior can go beyond the mean field approximation and yields good uncertainty on the performed experiments. Analysis on toy tasks shows that it can learn from significantly different tasks while finding similarities among them. Experiments of Mini-Imagenet yields the new state of the art with 74.5% accuracy on 5 shot learning. Finally, we provide experiments showing that other existing methods can fail to perform well in different benchmarks.
References in corpus (16)
- Overcoming catastrophic forgetting in neural networks
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Practical Bayesian Optimization of Machine Learning Algorithms
- Prototypical Networks for Few-shot Learning
- Weight Uncertainty in Neural Networks
- Matching Networks for One Shot Learning
- A Simple Neural Attentive Meta-Learner
- Deep Bayesian Active Learning with Image Data
- VIME: Variational Information Maximizing Exploration
- Modulating early visual processing by language
- Improving Variational Inference with Inverse Autoregressive Flow
- Recasting Gradient-Based Meta-Learning as Hierarchical Bayes
- Towards a Neural Statistician
- Multi-Level Variational Autoencoder: Learning Disentangled Representations from Grouped Observations
- Bayesian Multitask Learning with Latent Hierarchies
- Discriminative k-shot learning using probabilistic models
Cited by in corpus (7)
- A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
- Meta-Learning with Latent Embedding Optimization
- Meta-Learning with Warped Gradient Descent
- The Internet of Federated Things (IoFT): A Vision for the Future and In-depth Survey of Data-driven Approaches for Federated Learning
- Torchmeta: A Meta-Learning library for PyTorch
- Information Theoretic Meta Learning with Gaussian Processes
- A Few-Shot Sequential Approach for Object Counting