Active Boundary Annotation using Random MAP Perturbations

Subhransu Maji, Tamir Hazan, Tommi Jaakkola
Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, PMLR 33:604-613, 2014.

Abstract

We address the problem of efficiently annotating labels of objects when they are structured. Often the distribution over labels can be described using a joint potential function over the labels for which sampling is provably hard but efficient maximum a-posteriori (MAP) solvers exist. In this setting we develop novel entropy bounds that are based on the expected amount of perturbation to the potential function that is needed to change MAP decisions. By reasoning about the entropy reduction and cost tradeoff, our algorithm actively selects the next annotation task. As an example of our framework we propose a boundary refinement task which can used to obtain pixel-accurate image boundaries much faster than traditional tools by focussing on parts of the image for refinement in a multi-scale manner.

Cite this Paper

BibTeX
@InProceedings{pmlr-v33-maji14, title = {{Active Boundary Annotation using Random MAP Perturbations}}, author = {Maji, Subhransu and Hazan, Tamir and Jaakkola, Tommi}, booktitle = {Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics}, pages = {604--613}, year = {2014}, editor = {Kaski, Samuel and Corander, Jukka}, volume = {33}, series = {Proceedings of Machine Learning Research}, address = {Reykjavik, Iceland}, month = {22--25 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v33/maji14.pdf}, url = {https://proceedings.mlr.press/v33/maji14.html}, abstract = {We address the problem of efficiently annotating labels of objects when they are structured. Often the distribution over labels can be described using a joint potential function over the labels for which sampling is provably hard but efficient maximum a-posteriori (MAP) solvers exist. In this setting we develop novel entropy bounds that are based on the expected amount of perturbation to the potential function that is needed to change MAP decisions. By reasoning about the entropy reduction and cost tradeoff, our algorithm actively selects the next annotation task. As an example of our framework we propose a boundary refinement task which can used to obtain pixel-accurate image boundaries much faster than traditional tools by focussing on parts of the image for refinement in a multi-scale manner.} }
Endnote
%0 Conference Paper %T Active Boundary Annotation using Random MAP Perturbations %A Subhransu Maji %A Tamir Hazan %A Tommi Jaakkola %B Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2014 %E Samuel Kaski %E Jukka Corander %F pmlr-v33-maji14 %I PMLR %P 604--613 %U https://proceedings.mlr.press/v33/maji14.html %V 33 %X We address the problem of efficiently annotating labels of objects when they are structured. Often the distribution over labels can be described using a joint potential function over the labels for which sampling is provably hard but efficient maximum a-posteriori (MAP) solvers exist. In this setting we develop novel entropy bounds that are based on the expected amount of perturbation to the potential function that is needed to change MAP decisions. By reasoning about the entropy reduction and cost tradeoff, our algorithm actively selects the next annotation task. As an example of our framework we propose a boundary refinement task which can used to obtain pixel-accurate image boundaries much faster than traditional tools by focussing on parts of the image for refinement in a multi-scale manner.
RIS
TY - CPAPER TI - Active Boundary Annotation using Random MAP Perturbations AU - Subhransu Maji AU - Tamir Hazan AU - Tommi Jaakkola BT - Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics DA - 2014/04/02 ED - Samuel Kaski ED - Jukka Corander ID - pmlr-v33-maji14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 33 SP - 604 EP - 613 L1 - http://proceedings.mlr.press/v33/maji14.pdf UR - https://proceedings.mlr.press/v33/maji14.html AB - We address the problem of efficiently annotating labels of objects when they are structured. Often the distribution over labels can be described using a joint potential function over the labels for which sampling is provably hard but efficient maximum a-posteriori (MAP) solvers exist. In this setting we develop novel entropy bounds that are based on the expected amount of perturbation to the potential function that is needed to change MAP decisions. By reasoning about the entropy reduction and cost tradeoff, our algorithm actively selects the next annotation task. As an example of our framework we propose a boundary refinement task which can used to obtain pixel-accurate image boundaries much faster than traditional tools by focussing on parts of the image for refinement in a multi-scale manner. ER -
APA
Maji, S., Hazan, T. & Jaakkola, T.. (2014). Active Boundary Annotation using Random MAP Perturbations. Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 33:604-613 Available from https://proceedings.mlr.press/v33/maji14.html.

Related Material