From Instance-level Constraints to Space-Level Constraints: Making the Most of Prior Knowledge in Data Clustering

By Daniel Klein, Sepandar D. Kamvar, Christopher D. Manning. International Conference on Machine Learning, 2002

623 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/51556ef5eb3e1e18540177ce0201a047537d673d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Correlation clustering in general weighted graphs (2006, background): “There has been a great deal of work on solving clustering problems subject to constraints [18,29,30,15].”
CECM: Constrained evidential C-means algorithm (2012, background): “They have been incorporated in many different ways, generally by including a penalty term in the objective function [2, 13] or by altering the distances between objects with respect to the constraints [16, 30].”
Learning a Mahalanobis Metric from Equivalence Constraints (2005, methodology): “A different scenario, in which equivalence constraints are the naturalsource of training data, occurs when we wish to learn from several teachers who do not know each other and who are not able to coordinate among themselves the use of common labels.”
Active co-analysis of a set of shapes (2012, methodology): “These constraints have been incorporated into many clustering algorithms, such as K-means clustering [Wagstaff and Cardie 2000], hierarchical clustering [Klein et al. 2002], and clustering with Gaussian mixture models [Shental et al.”
Integrating constraints and metric learning in semi-supervised clustering (2004, methodology): “Several distance measures have been used for metric-based semi-supervised clustering including Euclidean distance trained by a shortest-path algorithm (Klein et al., 2002), string-edit distance learned using Expectation Maximization (EM) (Bilenko & Mooney, 2003), KL divergence adapted using gradient”
Clustering with Constraints: Feasibility Issues and the k-Means Algorithm (2005, methodology): “We can divide previous work on clustering under constraints into two types: 1) Where the constraints help the algorithm learn a distortion/distance/objective function [4, 15] and 2) Where the constraints are used as “hints” to guide the algorithm to a useful solution [18, 19].”
Accurate Annotation of Remote Sensing Images via Active Spectral Clustering with Little Expert Knowledge (2015, background): “Most studies on active clustering have built upon traditional clustering methods, such as k-means [36,41,42] and hierarchical clustering [27,37,43].”
Semi-supervised Kernel Metric Learning Using Relative Comparisons (2016, methodology): “The idea of semi-supervised clustering was initially introduced by Wagstaff and Cardie [2000], and since then a large number of different problem variants and methods have been proposed, the ﬁrst ones being COP-Kmeans [Wagstaff et al., 2001] and CCL [Klein et al., 2002].”
Semi-supervised graph clustering: a kernel approach (2005, background): “However, most semi-supervised clustering with pairwise constraints assumes that the input is in the form of data vectors (Wagstaff et al. 2001; Sinkkonen and Kaski 2002; Klein et al. 2002; Xing et al. 2003; Bar-Hillel et al. 2003; Basu et al. 2004b; Bilenko et al. 2004).”
Deep Network Embedding for Graph Representation Learning in Signed Networks (2019, methodology): “[27] proposed a semisupervised clustering algorithm by incorporating the”