Abstract:Knots are deeply entangled with every branch of science. One of the biggest open challenges in knot theory is to formalise a knot invariant that can unambiguously and efficiently distinguish any two knotted curves. Additionally, the conjecture that the geometrical embedding of a curve encodes information on its underlying topology is, albeit physically intuitive, far from proven. Here we attempt to tackle both these outstanding challenges by proposing a neural network (NN) approach that takes as input a geometric representation of a knotted curve and tries to make predictions of the curve's topology. Intriguingly, we discover that NNs trained with a so-called geometrical "local writhe" representation of a knot can distinguish curves that share one or many topological invariants and knot polynomials, such as mutant and composite knots, and can thus classify knotted curves more precisely than some knot polynomials. Additionally, we also show that our approach can be scaled up to classify all prime knots up to 10-crossings with more than 95\% accuracy. Finally, we show that our NNs can also be trained to solve knot localisation problems on open and closed curves. Our main discovery is that the pattern of "local writhe" is a potentially unique geometric signature of the underlying topology of a curve. We hope that our results will suggest new methods for quantifying generic entanglements in soft matter and even inform new topological invariants.

Knot Logic: Logical Connection and Topological Connection

Non-Reidemeister Knot Theory and Its Applications in Dynamical Systems, Geometry, and Topology

Hard Unknots and Collapsing Tangles

On Free Knots and Links

Graph-Based Logic and Sketches 1: The General Framework

On the Classification of Rational Knots

Slipknotting in Random Diagrams

Generalised knotoids

Nematics, Knots and Non-orientable Surfaces

Logical Patterns in Space

Compositions of Knots Using Alexander Polynomial

Links and dynamics

Geometric contexts and applications to logic

Loops and Knots as Topoi of Substance. Spinoza Revisited

Geometric Learning of Knot Topology

Topological Drawings meet Classical Theorems from Convex Geometry

Fregean Flows

Convex Embeddability and Knot Theory

Knot Theory for Proteins: Gauss Codes, Quandles and Bondles

Alpha-Structures and Ladders in Logical Geometry

Spatial logics with connectedness predicates