We use reinforcement learning as a means of constructing string compactifications with prescribed properties. Specifically, we study heterotic SOfalse(10false)$SO(10)$ GUT models on Calabi‐Yau three‐folds with monad bundles, in search of phenomenologically promising examples. Due to the vast number of bundles and the sparseness of viable choices, methods based on systematic scanning are not suitable for this class of models. By focusing on two specific manifolds with Picard numbers two and three, we show that reinforcement learning can be used successfully to explore monad bundles. Training can be accomplished with minimal computing resources and leads to highly efficient policy networks. They produce phenomenologically promising states for nearly 100% of episodes and within a small number of steps. In this way, hundreds of new candidate standard models are found.

Large classes of standard single‐field slow‐roll inflationary models consistent with the required number of e‐folds, the current bounds on the spectral index of scalar perturbations, the tensor‐to‐scalar ratio, and the scale of inflation can be efficiently constructed using genetic algorithms. The setup is modular and can be easily adapted to include further phenomenological constraints. A semi‐comprehensive search for sextic polynomial potentials results in ∼scriptOfalse(300,000false)$\sim \mathcal {O}(300,000)$ viable models for inflation. The analysis of this dataset reveals a preference for models with a tensor‐to‐scalar ratio in the range 0.0001≤r≤0.0004$0.0001\le r\le 0.0004$. We also consider potentials that involve cosine and exponential terms. In the last part we explore more complex methods of search relying on reinforcement learning and genetic programming. While reinforcement learning proves more difficult to use in this context, the genetic programming approach has the potential to uncover a multitude of viable inflationary models with new functional forms.

The immensity of the string landscape and the difficulty of identifying solutions that match the observed features of particle physics have raised serious questions about the predictive power of string theory. Modern methods of optimisation and search can, however, significantly improve the prospects of constructing the standard model in string theory. In this paper we scrutinise a corner of the heterotic string landscape consisting of compactifications on Calabi‐Yau three‐folds with monad bundles and show that genetic algorithms can be successfully used to generate anomaly‐free supersymmetric SOfalse(10false)$SO(10)$ GUTs with three families of fermions that have the right ingredients to accommodate the standard model. We compare this method with reinforcement learning and find that the two methods have similar efficacy but somewhat complementary characteristics.

We investigate reinforcement learning and genetic algorithms in the context of heterotic Calabi-Yau models with monad bundles. Both methods are found to be highly efficient in identifying phenomenologically attractive three-family models, in cases where systematic scans are not feasible. For monads on the bi-cubic Calabi-Yau either method facilitates a complete search of the environment and leads to similar sets of previously unknown three-family models.

In this paper, we apply reinforcement learning to the problem of constructing models in particle physics. As an example environment, we use the space of Froggatt-Nielsen type models for quark masses. Using a basic policy-based algorithm we show that neural networks can be successfully trained to construct Froggatt-Nielsen models which are consistent with the observed quark masses and mixing. The trained policy networks lead from random to phenomenologically acceptable models for over 90% of episodes and after an average episode length of about 20 steps. We also show that the networks are capable of finding models proposed in the literature when starting at nearby configurations.