Discriminative Embeddings of Latent Variable Models for Structured Data
This addresses scalability issues for researchers and practitioners in fields like computational biology and drug design, representing an incremental improvement over existing kernel methods.
The paper tackled the scalability and discriminative limitations of kernel methods for structured data by proposing structure2vec, an embedding approach that runs 2 times faster, produces models 10,000 times smaller, and achieves state-of-the-art predictive performance on millions of data points.
Kernel classifiers and regressors designed for structured data, such as sequences, trees and graphs, have significantly advanced a number of interdisciplinary areas such as computational biology and drug design. Typically, kernels are designed beforehand for a data type which either exploit statistics of the structures or make use of probabilistic generative models, and then a discriminative classifier is learned based on the kernels via convex optimization. However, such an elegant two-stage approach also limited kernel methods from scaling up to millions of data points, and exploiting discriminative information to learn feature representations. We propose, structure2vec, an effective and scalable approach for structured data representation based on the idea of embedding latent variable models into feature spaces, and learning such feature spaces using discriminative information. Interestingly, structure2vec extracts features by performing a sequence of function mappings in a way similar to graphical model inference procedures, such as mean field and belief propagation. In applications involving millions of data points, we showed that structure2vec runs 2 times faster, produces models which are $10,000$ times smaller, while at the same time achieving the state-of-the-art predictive performance.