Filtering with Abstract Particles

Jacob Steinhardt, Percy Liang
Proceedings of the 31st International Conference on Machine Learning, PMLR 32(1):727-735, 2014.

Abstract

Using particles, beam search and sequential Monte Carlo can approximate distributions in an extremely flexible manner. However, they can suffer from sparsity and inadequate coverage on large state spaces. We present a new filtering method that addresses this issue by using “abstract particles” that each represent an entire region of the state space. These abstract particles are combined into a hierarchical decomposition, yielding a representation that is both compact and flexible. Empirically, our method outperforms beam search and sequential Monte Carlo on both a text reconstruction task and a multiple object tracking task.

Cite this Paper

BibTeX
@InProceedings{pmlr-v32-steinhardt14, title = {Filtering with Abstract Particles}, author = {Steinhardt, Jacob and Liang, Percy}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {727--735}, year = {2014}, editor = {Xing, Eric P. and Jebara, Tony}, volume = {32}, number = {1}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/steinhardt14.pdf}, url = {https://proceedings.mlr.press/v32/steinhardt14.html}, abstract = {Using particles, beam search and sequential Monte Carlo can approximate distributions in an extremely flexible manner. However, they can suffer from sparsity and inadequate coverage on large state spaces. We present a new filtering method that addresses this issue by using “abstract particles” that each represent an entire region of the state space. These abstract particles are combined into a hierarchical decomposition, yielding a representation that is both compact and flexible. Empirically, our method outperforms beam search and sequential Monte Carlo on both a text reconstruction task and a multiple object tracking task.} }
Endnote
%0 Conference Paper %T Filtering with Abstract Particles %A Jacob Steinhardt %A Percy Liang %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-steinhardt14 %I PMLR %P 727--735 %U https://proceedings.mlr.press/v32/steinhardt14.html %V 32 %N 1 %X Using particles, beam search and sequential Monte Carlo can approximate distributions in an extremely flexible manner. However, they can suffer from sparsity and inadequate coverage on large state spaces. We present a new filtering method that addresses this issue by using “abstract particles” that each represent an entire region of the state space. These abstract particles are combined into a hierarchical decomposition, yielding a representation that is both compact and flexible. Empirically, our method outperforms beam search and sequential Monte Carlo on both a text reconstruction task and a multiple object tracking task.
RIS
TY - CPAPER TI - Filtering with Abstract Particles AU - Jacob Steinhardt AU - Percy Liang BT - Proceedings of the 31st International Conference on Machine Learning DA - 2014/01/27 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-steinhardt14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 32 IS - 1 SP - 727 EP - 735 L1 - http://proceedings.mlr.press/v32/steinhardt14.pdf UR - https://proceedings.mlr.press/v32/steinhardt14.html AB - Using particles, beam search and sequential Monte Carlo can approximate distributions in an extremely flexible manner. However, they can suffer from sparsity and inadequate coverage on large state spaces. We present a new filtering method that addresses this issue by using “abstract particles” that each represent an entire region of the state space. These abstract particles are combined into a hierarchical decomposition, yielding a representation that is both compact and flexible. Empirically, our method outperforms beam search and sequential Monte Carlo on both a text reconstruction task and a multiple object tracking task. ER -
APA
Steinhardt, J. & Liang, P.. (2014). Filtering with Abstract Particles. Proceedings of the 31st International Conference on Machine Learning, in Proceedings of Machine Learning Research 32(1):727-735 Available from https://proceedings.mlr.press/v32/steinhardt14.html.

Related Material