Learning collective variables that respect permutational symmetry

We propose algorithms for learning collective variables that respect permutational symmetry, with case studies on Lennard-Jones clusters using quantitative coarse-graining theory and tools from geometric data science.