Search

Liouville theory describes the dynamics of surfaces with constant negative curvature and can be used to study the Weil-Petersson geometry of the moduli space of Riemann surfaces. This leads to an efficient algorithm to compute the Weil--Petersson metric to arbitrary accuracy using Zamolodchikov's recursion relation for conformal blocks. For example, we compute the metric on $\mathcal M_{0,4}$ numerically to high accuracy by considering Liouville theory on a sphere with four punctures. We numerically compute the eigenvalues of the Weil-Petersson Laplacian, and find evidence that the obey the statistics of a random matrix in the Gaussian Orthogonal Ensemble., Comment: 29+6 pages

Large language models with a huge number of parameters, when trained on near internet-sized number of tokens, have been empirically shown to obey neural scaling laws: specifically, their performance behaves predictably as a power law in either parameters or dataset size until bottlenecked by the other resource. To understand this better, we first identify the necessary properties allowing such scaling laws to arise and then propose a statistical model -- a joint generative data model and random feature model -- that captures this neural scaling phenomenology. By solving this model in the dual limit of large training set size and large number of parameters, we gain insight into (i) the statistical structure of datasets and tasks that lead to scaling laws, (ii) the way nonlinear feature maps, such as those provided by neural networks, enable scaling laws when trained on these datasets, (iii) the optimality of the equiparameterization scaling of training sets and parameters, and (iv) whether such scaling laws can break down and how they behave when they do. Key findings are the manner in which the power laws that occur in the statistics of natural datasets are extended by nonlinear random feature maps and then translated into power-law scalings of the test loss and how the finite extent of the data's spectral power law causes the model's performance to plateau., Comment: 73 + 23 pages, 14 + 5 figures

We study symmetric products of the chiral 'Monster' conformal field theory with c=24 in the grand canonical ensemble by introducing a complex parameter \rho, whose imaginary part represents the chemical potential \mu conjugate to the number of copies of the CFT. The grand canonical partition function is given by the DMVV product formula in terms of the multiplicities of the seed CFT. It possesses an O(2,2;\ZZ) symmetry that enhances the familiar SL(2,\ZZ) modular invariance of the canonical ensemble and mixes the modular parameter \tau with the parameter \rho. By exploiting this enhanced modular symmetry and the pole structure of the DMVV product formula we are able to extend the region of validity of Cardy's formula, and explain why it matches the semi-classical Bekenstein-Hawking formula for black holes all the way down to the AdS-scale. We prove that for large c the spectrum contains a universal perturbative sector whose degeneracies obey Hagedorn growth. The transition from Cardy to Hagedorn growth is found to be due to the formation of a Bose-Einstein condensate of ground state CFTs at low temperatures. The grand canonical partition function has an interesting phase structure, which in addition to the standard Hawking-Page transition between low and high temperature, exhibits a wall-crossing transition that exchanges the roles of \tau and \rho., Comment: 28 pages, 3 figures