Multi-class Classification without Multi-class Labels
arXiv:1901.00544
Abstract
This work presents a new strategy for multi-class classification that requires no class-specific labels, but instead leverages pairwise similarity between examples, which is a weaker form of annotation. The proposed method, meta classification learning, optimizes a binary classifier for pairwise similarity prediction and through this process learns a multi-class classifier as a submodule. We formulate this approach, present a probabilistic graphical model for it, and derive a surprisingly simple loss function that can be used to learn neural network-based models. We then demonstrate that this same framework generalizes to the supervised, unsupervised cross-task, and semi-supervised settings. Our method is evaluated against state of the art in all three learning paradigms and shows a superior or comparable accuracy, providing evidence that learning multi-class classification without multi-class labels is a viable learning option.
International Conference on Learning Representations (ICLR 2019)
References in corpus (11)
- Adam: A Method for Stochastic Optimization
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
- Matching Networks for One Shot Learning
- Visualizing the Loss Landscape of Neural Nets
- On Constrained Spectral Clustering and Its Applications
- Realistic Evaluation of Deep Semi-Supervised Learning Algorithms
- Qualitatively characterizing neural network optimization problems
- Neural network-based clustering using pairwise constraints
- An empirical analysis of the optimization of deep network loss surfaces
- Semi-supervised Kernel Metric Learning Using Relative Comparisons