Abstract by Andrew Carr
NMT exploration of knots and their invariants
The study of knots in geometric topology has progressed recently with aid from computational methods. In this work, we explore the application of neural machine translation (NMT) to the generation of novel knots and studies of their distributions. Knots can be uniquely represented, up to reflection, by Dowker-Thistlethwaite (DT) codes. Additionally, knots have various invariants which can be derived from their projections. One of these invariants is the Jones polynomial. We use a novel representation of the Jones polynomial and a modern NMT architecture to explore transformations from knot invariants to the unique DT codes.