Learning Expanding Graphs for Signal Interpolation
This addresses cold start recommendation problems where graphs grow incrementally, but the approach appears incremental as it builds on existing graph signal processing methods.
The paper tackles the problem of signal interpolation on graphs that expand over time with unknown connectivity of new nodes, proposing a stochastic attachment model with data-driven parameter estimation. Numerical experiments on synthetic and real collaborative filtering data show the method's effectiveness, though no specific performance numbers are provided.
Performing signal processing over graphs requires knowledge of the underlying fixed topology. However, graphs often grow in size with new nodes appearing over time, whose connectivity is typically unknown; hence, making more challenging the downstream tasks in applications like cold start recommendation. We address such a challenge for signal interpolation at the incoming nodes blind to the topological connectivity of the specific node. Specifically, we propose a stochastic attachment model for incoming nodes parameterized by the attachment probabilities and edge weights. We estimate these parameters in a data-driven fashion by relying only on the attachment behaviour of earlier incoming nodes with the goal of interpolating the signal value. We study the non-convexity of the problem at hand, derive conditions when it can be marginally convexified, and propose an alternating projected descent approach between estimating the attachment probabilities and the edge weights. Numerical experiments with synthetic and real data dealing in cold start collaborative filtering corroborate our findings.