# Neural network-based clustering using pairwise constraints

arXiv: https://arxiv.org/abs/1511.06321 · HTML: https://mlsanity.com/paper/1511.06321/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Nov 2015.

## Built on by (most cited first, 22 of 22)

- [CLIP-GCD: Simple Language Guided Generalized Category Discovery](https://mlsanity.com/paper/2305.10420/) (2023)
- [Instance Level Affinity-Based Transfer for Unsupervised Domain Adaptation](https://mlsanity.com/paper/2104.01286/) (2021)
- [SLIC: Self-Supervised Learning with Iterative Clustering for Human Action Videos](https://mlsanity.com/paper/2206.12534/) (2022)
- [A Framework for Deep Constrained Clustering](https://mlsanity.com/paper/2101.02792/) (2021)
- [Deep Conditional Gaussian Mixture Model for Constrained Clustering](https://mlsanity.com/paper/2106.06385/) (2021)
- [Unseen Object Instance Segmentation with Fully Test-time RGB-D Embeddings Adaptation](https://mlsanity.com/paper/2204.09847/) (2022)
- [ADAIL: Adaptive Adversarial Imitation Learning](https://mlsanity.com/paper/2008.12647/) (2020)
- [Deep Clustering with Incomplete Noisy Pairwise Annotations: A Geometric Regularization Approach](https://mlsanity.com/paper/2305.19391/) (2023)
- [Cequel: Cost-Effective Querying of Large Language Models for Text Clustering](https://mlsanity.com/paper/2504.15640/) (2025)
- [Constrained Clustering and Multiple Kernel Learning without Pairwise Constraint Relaxation](https://mlsanity.com/paper/2203.12546/) (2022)
- [HPSCAN: Human Perception-Based Scattered Data Clustering](https://mlsanity.com/paper/2304.14185/) (2023)
- [ConstraintMatch for Semi-constrained Clustering](https://mlsanity.com/paper/2311.15395/) (2023)
- [Angular Constraint Embedding via SpherePair Loss for Constrained Clustering](https://mlsanity.com/paper/2510.06907/) (2025)
- [DMS: Differentiable Mean Shift for Dataset Agnostic Task Specific Clustering Using Side Information](https://mlsanity.com/paper/2305.18492/) (2023)
- [Enhancing Martian Terrain Recognition with Deep Constrained Clustering](https://mlsanity.com/paper/2503.17633/) (2025)
- [Exclusive Topic Modeling](https://mlsanity.com/paper/2102.03525/) (2021)
- [Exploring dual information in distance metric learning for clustering](https://mlsanity.com/paper/2105.12703/) (2021)
- [A Bibliographic View on Constrained Clustering](https://mlsanity.com/paper/2209.11125/) (2022)
- [Towards Balanced RGB-TSDF Fusion for Consistent Semantic Scene Completion by 3D RGB Feature Completion and a Classwise Entropy Loss Function](https://mlsanity.com/paper/2403.16888/) (2024)
- [Rethinking Open-World Semi-Supervised Learning: Distribution Mismatch and Inductive Inference](https://mlsanity.com/paper/2405.20829/) (2024)
- [Random Normed k-Means: A Paradigm-Shift in Clustering within Probabilistic Metric Spaces](https://mlsanity.com/paper/2504.03928/) (2025)
- [Amortized Neural Clustering of Time Series based on Statistical Features](https://mlsanity.com/paper/2605.13128/) (2026)

## Abstract (arXiv)

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.