MILEAGE: Multiple Instance LEArning with Global Embedding

Dan Zhang, Jingrui He, Luo Si, Richard Lawrence
Proceedings of the 30th International Conference on Machine Learning, PMLR 28(3):82-90, 2013.

Abstract

Multiple Instance Learning (MIL) methods generally represent each example as a collection of instances such that the features for local objects can be better captured, whereas traditional learning methods typically extract a global feature vector for each example as an integral part. However, there is limited research work on which of the two learning scenarios performs better. This paper proposes a novel framework – \emphMultiple Instance LEArning with Global Embedding (MILEAGE), in which the global feature vectors for traditional learning methods are integrated into the MIL setting. MILEAGE can leverage the benefits derived from both learning settings. Within the proposed framework, a large margin method is formulated. In particular, the proposed method adaptively tunes the weights on the two different kinds of feature representations (i.e., global and multiple instance) for each example and trains the classifier simultaneously. An alternative algorithm is proposed to solve the resulting optimization problem, which extends the bundle method to the non-convex case. Some important properties of the proposed method, such as the convergence rate and the generalization error rate, are analyzed. A series of experiments have been conducted to demonstrate the advantages of the proposed method over several state-of-the-art multiple instance and traditional learning methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v28-zhang13a, title = {MILEAGE: Multiple Instance LEArning with Global Embedding}, author = {Zhang, Dan and He, Jingrui and Si, Luo and Lawrence, Richard}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {82--90}, year = {2013}, editor = {Dasgupta, Sanjoy and McAllester, David}, volume = {28}, number = {3}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/zhang13a.pdf}, url = {https://proceedings.mlr.press/v28/zhang13a.html}, abstract = {Multiple Instance Learning (MIL) methods generally represent each example as a collection of instances such that the features for local objects can be better captured, whereas traditional learning methods typically extract a global feature vector for each example as an integral part. However, there is limited research work on which of the two learning scenarios performs better. This paper proposes a novel framework – \emphMultiple Instance LEArning with Global Embedding (MILEAGE), in which the global feature vectors for traditional learning methods are integrated into the MIL setting. MILEAGE can leverage the benefits derived from both learning settings. Within the proposed framework, a large margin method is formulated. In particular, the proposed method adaptively tunes the weights on the two different kinds of feature representations (i.e., global and multiple instance) for each example and trains the classifier simultaneously. An alternative algorithm is proposed to solve the resulting optimization problem, which extends the bundle method to the non-convex case. Some important properties of the proposed method, such as the convergence rate and the generalization error rate, are analyzed. A series of experiments have been conducted to demonstrate the advantages of the proposed method over several state-of-the-art multiple instance and traditional learning methods.} }
Endnote
%0 Conference Paper %T MILEAGE: Multiple Instance LEArning with Global Embedding %A Dan Zhang %A Jingrui He %A Luo Si %A Richard Lawrence %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-zhang13a %I PMLR %P 82--90 %U https://proceedings.mlr.press/v28/zhang13a.html %V 28 %N 3 %X Multiple Instance Learning (MIL) methods generally represent each example as a collection of instances such that the features for local objects can be better captured, whereas traditional learning methods typically extract a global feature vector for each example as an integral part. However, there is limited research work on which of the two learning scenarios performs better. This paper proposes a novel framework – \emphMultiple Instance LEArning with Global Embedding (MILEAGE), in which the global feature vectors for traditional learning methods are integrated into the MIL setting. MILEAGE can leverage the benefits derived from both learning settings. Within the proposed framework, a large margin method is formulated. In particular, the proposed method adaptively tunes the weights on the two different kinds of feature representations (i.e., global and multiple instance) for each example and trains the classifier simultaneously. An alternative algorithm is proposed to solve the resulting optimization problem, which extends the bundle method to the non-convex case. Some important properties of the proposed method, such as the convergence rate and the generalization error rate, are analyzed. A series of experiments have been conducted to demonstrate the advantages of the proposed method over several state-of-the-art multiple instance and traditional learning methods.
RIS
TY - CPAPER TI - MILEAGE: Multiple Instance LEArning with Global Embedding AU - Dan Zhang AU - Jingrui He AU - Luo Si AU - Richard Lawrence BT - Proceedings of the 30th International Conference on Machine Learning DA - 2013/05/26 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-zhang13a PB - PMLR DP - Proceedings of Machine Learning Research VL - 28 IS - 3 SP - 82 EP - 90 L1 - http://proceedings.mlr.press/v28/zhang13a.pdf UR - https://proceedings.mlr.press/v28/zhang13a.html AB - Multiple Instance Learning (MIL) methods generally represent each example as a collection of instances such that the features for local objects can be better captured, whereas traditional learning methods typically extract a global feature vector for each example as an integral part. However, there is limited research work on which of the two learning scenarios performs better. This paper proposes a novel framework – \emphMultiple Instance LEArning with Global Embedding (MILEAGE), in which the global feature vectors for traditional learning methods are integrated into the MIL setting. MILEAGE can leverage the benefits derived from both learning settings. Within the proposed framework, a large margin method is formulated. In particular, the proposed method adaptively tunes the weights on the two different kinds of feature representations (i.e., global and multiple instance) for each example and trains the classifier simultaneously. An alternative algorithm is proposed to solve the resulting optimization problem, which extends the bundle method to the non-convex case. Some important properties of the proposed method, such as the convergence rate and the generalization error rate, are analyzed. A series of experiments have been conducted to demonstrate the advantages of the proposed method over several state-of-the-art multiple instance and traditional learning methods. ER -
APA
Zhang, D., He, J., Si, L. & Lawrence, R.. (2013). MILEAGE: Multiple Instance LEArning with Global Embedding. Proceedings of the 30th International Conference on Machine Learning, in Proceedings of Machine Learning Research 28(3):82-90 Available from https://proceedings.mlr.press/v28/zhang13a.html.

Related Material