S. Chen, J. Xu, и T. Joachims. Proceedings of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, стр. 865--873. New York, NY, USA, ACM, (2013)
DOI: 10.1145/2487575.2487632
Аннотация
Learning algorithms that embed objects into Euclidean space have become the methods of choice for a wide range of problems, ranging from recommendation and image search to playlist prediction and language modeling. Probabilistic embedding methods provide elegant approaches to these problems, but can be expensive to train and store as a large monolithic model. In this paper, we propose a method that trains not one monolithic model, but multiple local embeddings for a class of pairwise conditional models especially suited for sequence and co-occurrence modeling. We show that computation and memory for training these multi-space models can be efficiently parallelized over many nodes of a cluster. Focusing on sequence modeling for music playlists, we show that the method substantially speeds up training while maintaining high model quality.
%0 Conference Paper
%1 chen2013multispace
%A Chen, Shuo
%A Xu, Jiexun
%A Joachims, Thorsten
%B Proceedings of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
%C New York, NY, USA
%D 2013
%I ACM
%K diss embedding inthesis lme logistic markov model modelling music playlist playlists prediction recommendation sequence
%P 865--873
%R 10.1145/2487575.2487632
%T Multi-space Probabilistic Sequence Modeling
%U http://doi.acm.org/10.1145/2487575.2487632
%X Learning algorithms that embed objects into Euclidean space have become the methods of choice for a wide range of problems, ranging from recommendation and image search to playlist prediction and language modeling. Probabilistic embedding methods provide elegant approaches to these problems, but can be expensive to train and store as a large monolithic model. In this paper, we propose a method that trains not one monolithic model, but multiple local embeddings for a class of pairwise conditional models especially suited for sequence and co-occurrence modeling. We show that computation and memory for training these multi-space models can be efficiently parallelized over many nodes of a cluster. Focusing on sequence modeling for music playlists, we show that the method substantially speeds up training while maintaining high model quality.
%@ 978-1-4503-2174-7
@inproceedings{chen2013multispace,
abstract = {Learning algorithms that embed objects into Euclidean space have become the methods of choice for a wide range of problems, ranging from recommendation and image search to playlist prediction and language modeling. Probabilistic embedding methods provide elegant approaches to these problems, but can be expensive to train and store as a large monolithic model. In this paper, we propose a method that trains not one monolithic model, but multiple local embeddings for a class of pairwise conditional models especially suited for sequence and co-occurrence modeling. We show that computation and memory for training these multi-space models can be efficiently parallelized over many nodes of a cluster. Focusing on sequence modeling for music playlists, we show that the method substantially speeds up training while maintaining high model quality.},
acmid = {2487632},
added-at = {2017-01-18T09:19:04.000+0100},
address = {New York, NY, USA},
author = {Chen, Shuo and Xu, Jiexun and Joachims, Thorsten},
biburl = {https://www.bibsonomy.org/bibtex/2ff88358be492e919b87cbd3636ffc6d3/becker},
booktitle = {Proceedings of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
doi = {10.1145/2487575.2487632},
interhash = {3f9acca95fa5dd0baf52ca469f6ad04e},
intrahash = {ff88358be492e919b87cbd3636ffc6d3},
isbn = {978-1-4503-2174-7},
keywords = {diss embedding inthesis lme logistic markov model modelling music playlist playlists prediction recommendation sequence},
location = {Chicago, Illinois, USA},
numpages = {9},
pages = {865--873},
publisher = {ACM},
series = {KDD '13},
timestamp = {2017-01-18T09:20:52.000+0100},
title = {Multi-space Probabilistic Sequence Modeling},
url = {http://doi.acm.org/10.1145/2487575.2487632},
year = 2013
}