Neural network-based clustering using pairwise constraints

Abstract

This paper presents a neural network-based end-to-end clustering framework. We design a novel strategy to utilize the contrastive criteria for pushing data-forming clusters directly from raw data, in addition to learning a feature embedding suitable for such clustering. The network is trained with weak labels, specifically partial pairwise relationships between data instances. The cluster assignments and their probabilities are then obtained at the output layer by feed-forwarding the data. The framework has the interesting characteristic that no cluster centers need to be explicitly specified, thus the resulting cluster distribution is purely data-driven and no distance metrics need to be predefined. The experiments show that the proposed approach beats the conventional two-stage method (feature embedding with k-means) by a significant margin. It also compares favorably to the performance of the standard cross entropy loss for classification. Robustness analysis also shows that the method is largely insensitive to the number of clusters. Specifically, we show that the number of dominant clusters is close to the true number of clusters even when a large k is used for clustering.

ICLR 2016

References in corpus (7)

• End-to-End Training of Deep Visuomotor Policies
• Learning to Compare Image Patches via Convolutional Neural Networks
• Unsupervised Deep Embedding for Clustering Analysis
• Deep Learning with Nonparametric Clustering
• Metric Learning with Adaptive Density Discrimination
• Unsupervised Feature Learning from Temporal Data
• Clustering With Side Information: From a Probabilistic Model to a Deterministic Algorithm

Cited by in corpus (17)

• Identifying Strong Lenses with Unsupervised Machine Learning using Convolutional Autoencoder
• Multi-class Classification without Multi-class Labels
• Automatic Discovery of Novel Intents & Domains from Text Utterances
• Machine Learning for Cataract Classification and Grading on Ophthalmic Imaging Modalities: A Survey
• Deep Image Category Discovery using a Transferred Similarity Function
• A Framework for Deep Constrained Clustering -- Algorithms and Advances
• Clustering-Oriented Representation Learning with Attractive-Repulsive Loss
• Human-like Clustering with Deep Convolutional Neural Networks
• ADAIL: Adaptive Adversarial Imitation Learning
• Multi-Class Classification from Noisy-Similarity-Labeled Data
• Interpretable Structured Learning with Sparse Gated Sequence Encoder for Protein-Protein Interaction Prediction
• A probabilistic constrained clustering for transfer learning and image category discovery
• Local Clustering with Mean Teacher for Semi-supervised Learning
• Consistency Regularization with Generative Adversarial Networks for Semi-Supervised Learning
• A Framework for Deep Constrained Clustering
• Exclusive Topic Modeling
• Unique Identification of Macaques for Population Monitoring and Control