Equilibrium Distribution for t-Distributed Stochastic Neighbor Embedding with Generalized Kernels
The work extends t-SNE by introducing generalized input and output kernels to allow for a wider range of weighting functions. By imposing specific mathematical conditions on these kernels, the algorithm is proven to converge to a stable equilibrium distribution as the number of data points increases. This establishes theoretical guarantees for t-SNE under more flexible kernel definitions.