Automatic Tuning of Interactive Perception Applications

Qian Zhu, Branislav Kveton, Lily Mummert, Padmanabhan Pillai
Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, PMLR R8:742-750, 2010.

Abstract

Interactive applications incorporating high-data rate sensing and computer vision are becoming possible due to novel runtime systems and the use of parallel computation resources. To allow interactive use, such applications require care- ful tuning of multiple application parameters to meet required fidelity and latency bounds. This is a nontrivial task, often requiring expert knowl- edge, which becomes intractable as resources and application load characteristics change. This paper describes a method for automatic perfor- mance tuning that learns application character- istics and effects of tunable parameters online, and constructs models that are used to maximize fidelity for a given latency constraint. The pa- per shows that accurate latency models can be learned online, knowledge of application struc- ture can be used to reduce the complexity of the learning task, and operating points can be found that achieve 90% of the optimal fidelity by ex- ploring the parameter space only 3% of the time.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR8-zhu10a, title = {Automatic Tuning of Interactive Perception Applications}, author = {Zhu, Qian and Kveton, Branislav and Mummert, Lily and Pillai, Padmanabhan}, booktitle = {Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence}, pages = {742--750}, year = {2010}, editor = {Grünwald, Peter and Spirtes, Peter}, volume = {R8}, series = {Proceedings of Machine Learning Research}, month = {08--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r8/main/assets/zhu10a/zhu10a.pdf}, url = {https://proceedings.mlr.press/r8/zhu10a.html}, abstract = {Interactive applications incorporating high-data rate sensing and computer vision are becoming possible due to novel runtime systems and the use of parallel computation resources. To allow interactive use, such applications require care- ful tuning of multiple application parameters to meet required fidelity and latency bounds. This is a nontrivial task, often requiring expert knowl- edge, which becomes intractable as resources and application load characteristics change. This paper describes a method for automatic perfor- mance tuning that learns application character- istics and effects of tunable parameters online, and constructs models that are used to maximize fidelity for a given latency constraint. The pa- per shows that accurate latency models can be learned online, knowledge of application struc- ture can be used to reduce the complexity of the learning task, and operating points can be found that achieve 90% of the optimal fidelity by ex- ploring the parameter space only 3% of the time.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Automatic Tuning of Interactive Perception Applications %A Qian Zhu %A Branislav Kveton %A Lily Mummert %A Padmanabhan Pillai %B Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2010 %E Peter Grünwald %E Peter Spirtes %F pmlr-vR8-zhu10a %I PMLR %P 742--750 %U https://proceedings.mlr.press/r8/zhu10a.html %V R8 %X Interactive applications incorporating high-data rate sensing and computer vision are becoming possible due to novel runtime systems and the use of parallel computation resources. To allow interactive use, such applications require care- ful tuning of multiple application parameters to meet required fidelity and latency bounds. This is a nontrivial task, often requiring expert knowl- edge, which becomes intractable as resources and application load characteristics change. This paper describes a method for automatic perfor- mance tuning that learns application character- istics and effects of tunable parameters online, and constructs models that are used to maximize fidelity for a given latency constraint. The pa- per shows that accurate latency models can be learned online, knowledge of application struc- ture can be used to reduce the complexity of the learning task, and operating points can be found that achieve 90% of the optimal fidelity by ex- ploring the parameter space only 3% of the time. %Z Reissued by PMLR on 04 October 2026.
APA
Zhu, Q., Kveton, B., Mummert, L. & Pillai, P.. (2010). Automatic Tuning of Interactive Perception Applications. Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R8:742-750 Available from https://proceedings.mlr.press/r8/zhu10a.html. Reissued by PMLR on 04 October 2026.

Related Material