Our algorithm incorporates constraints into the Riemannian version of Hamiltonian Monte Carlo and maintains sparsity.

Projected 23] adds smaller .

You are right in that the proof makes direct use of existing9 linear programming results. At first glance the difference in our35 perspectives seems to be about how the optimization process discovers the representation, rather than the space of36 possible representations, but if you are making a more specific distinction please let us know in the revised review.37

Machine learning from explanations (MLX) is an approach to learning that uses human-provided explanations of relevant or irrelevant features for each input to ensure that model predictions are right for the right reasons . Existing MLX approaches rely on local model interpretation methods and require strong model smoothing to align model and human explanations, leading to sub-optimal performance.

Whenconsidering networkdata,regularconvolutionsdonot hold since the data domain presents an irregular structure given by the network topology.