EM-DD: An Improved Multiple-Instance Learning Technique

By Sally A. Goldman. Neural Information Processing Systems, 2001

761 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/fb98ba3aa910e0ea4bd16c755832a086dc2bb2a2

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

Revisiting Multiple-Instance Learning Via Embedded Instance Selection (2008, methodology): “Since then, many other MI learning algorithms have been proposed (see, for example, [1], [2], [10], [12], [14], [24], [25], [27], [28], [29], [30]).”
A review of multi-instance learning assumptions (2010, methodology): “A refinement of this algorithm, expectation maximization (EM)-diverse density (DD), was proposed by Zhang and Goldman (2001).”
MILES: Multiple-Instance Learning via Embedded Instance Selection (2006, methodology): “Zhang and Goldman [61] combined the idea of expectation-maximization (EM) with DD and developed an algorithm, EM-DD, to search for the most likely concept.”
Image Categorization by Learning and Reasoning with Regions (2004, methodology): “The EM-DD result cited in this paper is provide by Andrews et al. (2003) using the correct algorithm.”
A Multiple Instance Learning Strategy for Combating Good Word Attacks on Spam Filters (2008, methodology): “EM-DD combines EM with the DD algorithm to reduce the multiple instance learning problem to a single-instance setting.”
Multiple-instance ranking: Learning to rank images for image retrieval (2008, methodology): “This problem has been studied by a lot of researchers and many algorithms have been developed, such as Diverse Density [18], EM-DD [24], MI- SVM and mi-SVM [2].”
Multiple Instance Learning: Algorithms and Applications (2008, methodology): “The EM-DD [22] algorithm uses the same cost function as the above, but the optimization technique is different.”
Multiple instance learning: A survey of problem characteristics and applications (2016, methodology): “As discussed in Section 6.3, some methods are optimized for bag classification (EM-DD, MI-SVM, MILES, MILBoost, RSIS-EoSVM).”
Multiple instance ranking (2008, background): “Algorithms for MIC stem from diverse density(Maron&Ratan, 1998; Zhang & Goldman, 2001), neural networks (Ramon & Figure 2.”
Machine Learning and Radiology (2012, methodology): “This algorithm was improved by incorporating expectation maximization (EM) to estimate which instance(s) in a bag is responsible for the assigned class label (Zhang and Goldman, 2001).”
Multi-instance multi-label learning (2008, methodology): “To name a few, Diverse Density [45] and Em-dd [83], k-nearest neighbor algorithms Citation-knn and Bayesian-knn [67], decision tree algorithms Relic [54] and Miti [9], neural network algorithms Bp-mip and extensions [77,90] and Rbf-mip [78], rule learning algorithm Ripper-mi [20], support vector mac”
Weakly Supervised Action Detection (2011, methodology): “However, these methods either iteratively select positive instances locally [11, 19] (that is, each positive instance is selected independently of each other), or select the positive instances globally (by considering only the distances between positive and negative instances).”
Multiple instance learning with generalized support vector machines (2002, background): “, [8, 12]) have focused on specially tailored machine learning algorithms that do not compare favorably in the limiting case of bags of size 1 (the standard classification setting).”
Landmark-based deep multi-instance learning for brain disease diagnosis (2017, methodology): “Speciﬁcally, we ﬁrst segment the brain into three different tissue types, i.e. , gray matter (GM), white matter (WM), and cerebrospinal ﬂuid (CSF), using FAST ( Zhang et al., 2001 ) in the FSL software package.”
Multiple instance learning for soft bags via top instances (2015, unknown): “EM-DD [28] 84.”
Multi-Instance Kernels (2002, methodology): “While EM-DD (Zhang & Goldman, 2002) is superior on both datasets, among the algorithms that are based on general purpose learning methods, SVMs with MI kernels outperform all other methods.”