Learning to cluster in order to transfer across domains and tasks

Abstract

This paper introduces a novel method to perform transfer learning across domains and tasks, formulating it as a problem of learning to cluster. The key insight is that, in addition to features, we can transfer similarity information and this is sufficient to learn a similarity function and clustering network to perform both domain adaptation and cross-task transfer learning. We begin by reducing categorical information to pairwise constraints, which only considers whether two instances belong to the same class or not. This similarity is category-agnostic and can be learned from data in the source domain using a similarity network. We then present two novel approaches for performing transfer learning using this similarity function. First, for unsupervised domain adaptation, we design a new loss function to regularize classification with a constrained clustering loss, hence learning a clustering network with the transferred similarity metric generating the training inputs. Second, for cross-task learning (i.e., unsupervised clustering with unseen categories), we propose a framework to reconstruct and estimate the number of semantic clusters, again using the clustering network. Since the similarity network is noisy, the key is to use a robust clustering algorithm, and we show that our formulation is more robust than the alternative constrained and unconstrained clustering approaches. Using this method, we first show state of the art results for the challenging cross-task problem, applied on Omniglot and ImageNet. Our results show that we can reconstruct semantic clusters with high accuracy. We then evaluate the performance of cross-domain transfer using images from the Office-31 and SVHN-MNIST tasks and present top accuracy on both datasets. Our approach doesn't explicitly deal with domain discrepancy. If we combine with a domain adaptation loss, it shows further improvement.

ICLR 2018

References in corpus (5)

• Learning Transferable Features with Deep Adaptation Networks
• Deep Domain Confusion: Maximizing for Domain Invariance
• Asymmetric Tri-training for Unsupervised Domain Adaptation
• Adversarial Discriminative Domain Adaptation
• Learning to Compare Image Patches via Convolutional Neural Networks

Cited by in corpus (19)

• A Closer Look at Few-shot Classification
• Discovering New Intents with Deep Aligned Clustering
• Large-Scale Long-Tailed Recognition in an Open World
• Open-World Semi-Supervised Learning
• FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space
• TEXTOIR: An Integrated and Visualized Platform for Text Open Intent Recognition
• Learning to Discover Novel Visual Categories via Deep Transfer Clustering
• Temporal Attentive Alignment for Large-Scale Video Domain Adaptation
• Meta Discovery: Learning to Discover Novel Classes given Very Limited Data
• Are Few-Shot Learning Benchmarks too Simple ? Solving them without Task Supervision at Test-Time
• Weak-shot Fine-grained Classification via Similarity Transfer
• Detection of Paroxysmal Atrial Fibrillation using Attention-based Bidirectional Recurrent Neural Networks
• Learning from the Scene and Borrowing from the Rich: Tackling the Long Tail in Scene Graph Generation
• A probabilistic constrained clustering for transfer learning and image category discovery
• Discriminative Clustering for Robust Unsupervised Domain Adaptation
• Joint Representation Learning and Novel Category Discovery on Single- and Multi-modal Data
• Mining Entity Synonyms with Efficient Neural Set Generation
• Fair Meta-Learning For Few-Shot Classification
• Open-Set Representation Learning through Combinatorial Embedding