Input Space Mode Connectivity in Deep Neural Networks

J Vrabel, O Shem-Ur, Y Oz, D Krueger - arxiv preprint arxiv:2409.05800, 2024 - arxiv.org
We extend the concept of loss landscape mode connectivity to the input space of deep
neural networks. Mode connectivity was originally studied within parameter space, where it …

A singular Riemannian Geometry Approach to Deep Neural Networks III. Piecewise Differentiable Layers and Random Walks on -dimensional Classes

A Benfenati, A Marta - arxiv preprint arxiv:2404.06104, 2024 - arxiv.org
Neural networks are playing a crucial role in everyday life, with the most modern generative
models able to achieve impressive results. Nonetheless, their functioning is still not very …