NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:4485
Title:Modeling Dynamic Functional Connectivity with Latent Factor Gaussian Processes

The manuscript proposes a probabilistic model for dynamic functional connectivity. Novel components include enforcing positive definiteness using a log representation. The reviewers discuss the novelty of the modeling approach and the theoretical claims as strengths of the manuscript. On the other hand, the reviewers point out major weaknesses which are primarily empirical, including missing details on key hyper-parameters, missing comparison to standard baselines, and insufficient empirical validation i.e. using more than one real dataset. Overall, the reviewers and AC agree that this is a timely and interesting submission, which can be improved in terms of empirical evaluation.