Multi-task classification with sequential instances and tasks

Xu, Wei; Liu, Wei; Chi, Haoyuan; Huang, Xiaolin; Yang, Jie

doi:10.1016/j.image.2018.02.013

Cited by 5 publications

(1 citation statement)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, manually tuning these hyperparameters are expensive and intractable in web-scale recommendation scenarios. An alternative approach [23] in practice is to optimize each task iteratively, which, may stick at a local optimum and fail to perform well in some tasks. Instead, we use the idea of homoscedastic uncertainty [13] to weigh the loss of each task automatically during model training.…”

Section: Learning Task Weights With Homoscedastic Uncertaintymentioning

confidence: 99%

M2GRL: A Multi-task Multi-view Graph Representation Learning Framework for Web-scale Recommender Systems

Wang

Lin

et al. 2020

Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery &Amp; Data Mining

View full text Add to dashboard Cite

Combining graph representation learning with multi-view data (side information) for recommendation is a trend in industry. Most existing methods can be categorized as multi-view representation fusion; they first build one graph and then integrate multi-view data into a single compact representation for each node in the graph. However, these methods are raising concerns in both engineering and algorithm aspects: 1) multi-view data are abundant and informative in industry and may exceed the capacity of one single vector, and 2) inductive bias may be introduced as multi-view data are often from different distributions. In this paper, we use a multi-view representation alignment approach to address this issue. Particularly, we propose a multi-task multi-view graph representation learning framework (M2GRL) to learn node representations from multi-view graphs for web-scale recommender systems. M2GRL constructs one graph for each single-view data, learns multiple separate representations from multiple graphs, and performs alignment to model cross-view relations. M2GRL chooses a multi-task learning paradigm to learn intra-view representations and cross-view relations jointly. Besides, M2GRL applies homoscedastic uncertainty to adaptively tune the loss weights of tasks during training. We deploy M2GRL at Taobao and train it on 57 billion examples. According to offline metrics and online A/B tests, M2GRL significantly outperforms other state-of-the-art algorithms. Further exploration on diversity recommendation in Taobao shows the effectiveness of utilizing multiple representations produced by M2GRL, which we argue is a promising direction for various industrial recommendation tasks of different focus. A demo code of M2GRL is released at https://github.com/99731/M2GRL.

show abstract

Section: Learning Task Weights With Homoscedastic Uncertaintymentioning

confidence: 99%