Multi-Modal Bayesian Embeddings for Learning Social Knowledge Graphs
This work addresses the challenge of integrating social networks with knowledge bases for applications like academic search, though it appears incremental as it builds on existing embedding techniques.
The authors tackled the problem of connecting online social networks to open knowledge bases by proposing GenVector, a multi-modal Bayesian embedding model, which outperformed state-of-the-art methods on three datasets and significantly decreased error rates in an online A/B test on a large-scale academic system.
We study the extent to which online social networks can be connected to open knowledge bases. The problem is referred to as learning social knowledge graphs. We propose a multi-modal Bayesian embedding model, GenVector, to learn latent topics that generate word and network embeddings. GenVector leverages large-scale unlabeled data with embeddings and represents data of two modalities---i.e., social network users and knowledge concepts---in a shared latent topic space. Experiments on three datasets show that the proposed method clearly outperforms state-of-the-art methods. We then deploy the method on AMiner, a large-scale online academic search system with a network of 38,049,189 researchers with a knowledge base with 35,415,011 concepts. Our method significantly decreases the error rate in an online A/B test with live users.