DMCA
Cold-start Active Learning with Robust Ordinal Matrix Factorization
| Citations: | 1 - 0 self |
BibTeX
@MISC{Houlsby_cold-startactive,
author = {Neil Houlsby and Jose ́ Miguel Hernández-lobato and Zoubin Ghahramani},
title = {Cold-start Active Learning with Robust Ordinal Matrix Factorization},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We present a new matrix factorization model for rating data and a corresponding active learning strategy to address the cold-start problem. Cold-start is one of the most challenging tasks for rec-ommender systems: what to recommend with new users or items for which one has little or no data. An approach is to use active learning to collect the most useful initial ratings. How-ever, the performance of active learning depends strongly upon having accurate estimates of i) the uncertainty in model parameters and ii) the in-trinsic noisiness of the data. To achieve these estimates we propose a heteroskedastic Bayesian model for ordinal matrix factorization. We also present a computationally efficient framework for Bayesian active learning with this type of complex probabilistic model. This algorithm successfully distinguishes between informative and noisy data points. Our model yields state-of-the-art predictive performance and, coupled with our active learning strategy, enables us to gain useful information in the cold-start setting from the very first active sample. 1.
Keyphrases
robust ordinal matrix factorization cold-start active learning bayesian active learning rec-ommender system useful information ordinal matrix factorization active learning strategy noisy data point in-trinsic noisiness complex probabilistic model first active sample corresponding active learning strategy model parameter heteroskedastic bayesian model new matrix factorization model active learning model yield state-of-the-art predictive performance cold-start setting useful initial rating accurate estimate rating data cold-start problem challenging task efficient framework new user
