Multi-instance multi-label learning in the presence of novel class instances

Anh Pham, Raviv Raich, Xiaoli Fern, Jesús Pérez Arriaga
Proceedings of the 32nd International Conference on Machine Learning, PMLR 37:2427-2435, 2015.

Abstract

Multi-instance multi-label learning (MIML) is a framework for learning in the presence of label ambiguity. In MIML, experts provide labels for groups of instances (bags), instead of directly providing a label for every instance. When labeling efforts are focused on a set of target classes, instances outside this set will not be appropriately modeled. For example, ornithologists label bird audio recordings with a list of species present. Other additional sound instances, e.g., a rain drop or a moving vehicle sound, are not labeled. The challenge is due to the fact that for a given bag, the presence or absence of novel instances is latent. In this paper, this problem is addressed using a discriminative probabilistic model that accounts for novel instances. We propose an exact and efficient implementation of the maximum likelihood approach to determine the model parameters and consequently learn an instance-level classifier for all classes including the novel class. Experiments on both synthetic and real datasets illustrate the effectiveness of the proposed approach.

Cite this Paper

BibTeX
@InProceedings{pmlr-v37-pham15, title = {Multi-instance multi-label learning in the presence of novel class instances}, author = {Pham, Anh and Raich, Raviv and Fern, Xiaoli and Arriaga, Jesús Pérez}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning}, pages = {2427--2435}, year = {2015}, editor = {Bach, Francis and Blei, David}, volume = {37}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v37/pham15.pdf}, url = {https://proceedings.mlr.press/v37/pham15.html}, abstract = {Multi-instance multi-label learning (MIML) is a framework for learning in the presence of label ambiguity. In MIML, experts provide labels for groups of instances (bags), instead of directly providing a label for every instance. When labeling efforts are focused on a set of target classes, instances outside this set will not be appropriately modeled. For example, ornithologists label bird audio recordings with a list of species present. Other additional sound instances, e.g., a rain drop or a moving vehicle sound, are not labeled. The challenge is due to the fact that for a given bag, the presence or absence of novel instances is latent. In this paper, this problem is addressed using a discriminative probabilistic model that accounts for novel instances. We propose an exact and efficient implementation of the maximum likelihood approach to determine the model parameters and consequently learn an instance-level classifier for all classes including the novel class. Experiments on both synthetic and real datasets illustrate the effectiveness of the proposed approach.} }
Endnote
%0 Conference Paper %T Multi-instance multi-label learning in the presence of novel class instances %A Anh Pham %A Raviv Raich %A Xiaoli Fern %A Jesús Pérez Arriaga %B Proceedings of the 32nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Francis Bach %E David Blei %F pmlr-v37-pham15 %I PMLR %P 2427--2435 %U https://proceedings.mlr.press/v37/pham15.html %V 37 %X Multi-instance multi-label learning (MIML) is a framework for learning in the presence of label ambiguity. In MIML, experts provide labels for groups of instances (bags), instead of directly providing a label for every instance. When labeling efforts are focused on a set of target classes, instances outside this set will not be appropriately modeled. For example, ornithologists label bird audio recordings with a list of species present. Other additional sound instances, e.g., a rain drop or a moving vehicle sound, are not labeled. The challenge is due to the fact that for a given bag, the presence or absence of novel instances is latent. In this paper, this problem is addressed using a discriminative probabilistic model that accounts for novel instances. We propose an exact and efficient implementation of the maximum likelihood approach to determine the model parameters and consequently learn an instance-level classifier for all classes including the novel class. Experiments on both synthetic and real datasets illustrate the effectiveness of the proposed approach.
RIS
TY - CPAPER TI - Multi-instance multi-label learning in the presence of novel class instances AU - Anh Pham AU - Raviv Raich AU - Xiaoli Fern AU - Jesús Pérez Arriaga BT - Proceedings of the 32nd International Conference on Machine Learning DA - 2015/06/01 ED - Francis Bach ED - David Blei ID - pmlr-v37-pham15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 37 SP - 2427 EP - 2435 L1 - http://proceedings.mlr.press/v37/pham15.pdf UR - https://proceedings.mlr.press/v37/pham15.html AB - Multi-instance multi-label learning (MIML) is a framework for learning in the presence of label ambiguity. In MIML, experts provide labels for groups of instances (bags), instead of directly providing a label for every instance. When labeling efforts are focused on a set of target classes, instances outside this set will not be appropriately modeled. For example, ornithologists label bird audio recordings with a list of species present. Other additional sound instances, e.g., a rain drop or a moving vehicle sound, are not labeled. The challenge is due to the fact that for a given bag, the presence or absence of novel instances is latent. In this paper, this problem is addressed using a discriminative probabilistic model that accounts for novel instances. We propose an exact and efficient implementation of the maximum likelihood approach to determine the model parameters and consequently learn an instance-level classifier for all classes including the novel class. Experiments on both synthetic and real datasets illustrate the effectiveness of the proposed approach. ER -
APA
Pham, A., Raich, R., Fern, X. & Arriaga, J.P.. (2015). Multi-instance multi-label learning in the presence of novel class instances. Proceedings of the 32nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 37:2427-2435 Available from https://proceedings.mlr.press/v37/pham15.html.

Related Material