Forschungsleitung (Direction de recherche) : Patrick GALLINARI

Co-Betreuung : PIWOWARSKI Benjamin

Representation Learning for Relational Data

First, we proposeda model for heterogeneous graph node classification. Contrary to other state of the art models, our solution is able to learn edge weights when learning simultaneously the representations and the classifiers. To handle the incertainty of the entity representation, we have got interested in using gaussian representations. First of all, we extended our first model, then, we proposed an algorithm for relational time series forecasting where the observations are not only correlated inside each series, but also across the different series.
At last, we apply the Gaussian representation learning approach to the collaborative filtering task.
This is a preliminary work to see if the properties of Gaussian representations found on the two previous tasks were also verified for the ranking one.
The goal of this work was then to generalize the approach to more relational data and not only bipartite graphs between users and items.
Thus, we used gaussian representations for collaborative filtering, and showed their benefits from a theoretical and experimental point of view.

Publikationen 2015-2018