Uncertainty in Multitask Transfer Learning
This addresses the challenge of efficient knowledge transfer in multitask learning for AI applications, though it appears incremental as it builds on variational Bayes neural networks.
The paper tackles the problem of accumulating knowledge from multiple tasks to improve few-shot learning, achieving a new state-of-the-art accuracy of 74.5% on Mini-Imagenet for 5-shot learning.
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.