A review of multi-instance learning assumptions

By James R. Foulds, Eibe Frank. 2010

420 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Systems

Citations in the corpus

Learning Theory for Distribution Regression (2014, methodology): “y his/her time series trials, or a customer might be characterized by his/her shopping records. The MIL approach has been applied in several domains; see the reviews from Babenko (2004); Zhou (2004); Foulds and Frank (2010); Amores (2013). “Bag-of-objects” methods (MIL, classiﬁcation): Despite the l”
Monitoring Land-Cover Changes: A Machine-Learning Perspective (2016, background): “A survey of multiinstance learning approaches can be found in [53] and [54].”
Clustering for point pattern data (2017, background): “In multiple instance learning [22], [23], the ‘bags’ are indeed point patterns.”
From Group to Individual Labels Using Deep Features (2015, background): “…aggregation function over instance labels is an OR function, i.e., that positive bags contain at 1 The literature on this topic is vast and there is often disagreement in terminology—for extensive surveys see [5, 9] least 1 positive instance and that negative bags contain only negative instances.”
Not‐so‐supervised: A survey of semi‐supervised, multi‐instance, and transfer learning in medical image analysis (2018, methodology): “[2016] cell segmentation in microscopy graph-based, active Multiple Gass et al. [2012] segmentation in two applications graph-based Ciurte et al.”
Multi-Instance Dimensionality Reduction (2010, background): “There are alternative multi-instance learning assumptions that have been found useful in practice (Foulds and Frank 2009).”
Patch-based Convolutional Neural Network for Whole Slide Tissue Image Classification (2015, methodology): “Moreover, this method can be thought of as a Count-based Multiple Instance (CMI) learning method with two-level learning [49] which is a more general MIL assumption [20] than the Standard Multiple Instance (SMI) assumption.”
Multi-instance multi-label learning (2008, background): “As the research of multi-instance learning goes on, however, some other assumptions have been introduced [29].”
Modern Hopfield Networks and Attention for Immune Repertoire Classification (2020, background): “Note that this poses a credit assignment problem, since the sequences that are responsible for the label Y have to be identified and that the relation between instance-label and bag-label can be more complex (Foulds & Frank, 2010).”
Audio Set Classification with Attention Model: A Probabilistic Perspective (2017, background): “The weighted collective assumption is a generalization of the collective assumption, where a weight w(x) is allowed for each instance x [9]:”
Is object localization for free? - Weakly-supervised learning with convolutional neural networks (2015, positive): “The proposed method can also be seen as a variant of Multiple Instance Learning [21, 29, 54] if we refer to each image as a “bag” and treat each image window as a “sample”.”
Model-Based Multiple Instance Learning (2017, background): “The emerging area of machine learning dedicated to point pattern data is commonly known as Multiple Instance (MI) learning [1], [2].”
A brief introduction to weakly supervised learning (2018, neutral): “This is called multi-instance learning [28, 31].”
Model-based classification and novelty detection for point pattern data (2016, background): “However, the broader task of learning from point pattern data is more appropriately posed as a Multiple Instance Learning (MIL) problem [9], [10], since multiple instance data or ‘bags’ are indeed point patterns.”