Description
I will present a new approach to bootstrapping string-like theories using a one-parameter family of local, crossing symmetric dispersion relations and field-definition ambiguities. This enables us to use mass-level truncation and go beyond the dual resonance hypothesis. Remarkably, we find that imposing entanglement minimization along with some low-energy constraints leads to an excellent approximation to the superstring amplitudes. We also find other interesring S-matrices that do not obey the duality hypothesis, but exhibit a transition from Regge behaviour to power law behaviour at high energies. In addition to using SDPB to impose the unitarity constraints as is typical, we also impose non-linear constraints using a Physics-Informed Neural Network (PINN). This is the first bootstrap study that uses PINNs for non-linear, constrained optimization.
Based on: arXiv:2409.18259 [hep-th]
