Learning to simulate high energy particle collisions from unlabeled data.

In many scientific fields which rely on statistical inference, simulations are often used to map from theoretical models to experimental data, allowing scientists to test model predictions against experimental results. Experimental data is often reconstructed from indirect measurements causing the aggregate transformation from theoretical models to experimental data to be poorly-described analytically. Instead, numerical simulations are used at great computational cost. We introduce Optimal-Transport-based Unfolding and Simulation (OTUS), a fast simulator based on unsupervised machine-learning that is capable of predicting experimental data from theoretical models. Without the aid of current simulation information, OTUS trains a probabilistic autoencoder to transform directly between theoretical models and experimental data. Identifying the probabilistic autoencoder's latent space with the space of theoretical models causes the decoder network to become a fast, predictive simulator with the potential to replace current, computationally-costly simulators. Here, we provide proof-of-principle results on two particle physics examples, Z-boson and top-quark decays, but stress that OTUS can be widely applied to other fields.

Bounds on invisible Higgs boson decays extracted from LHC ttH production data.

We present an upper bound on the branching fraction of the Higgs boson to invisible particles by recasting a CMS Collaboration search for stop quarks decaying to tt + E(T)(miss). The observed (expected) bound, BF(H → inv.) < 0.40(0.65) at 95% C.L., is the strongest direct limit to date, benefiting from a downward fluctuation in the CMS data in that channel. In addition, we combine this new constraint with existing published constraints to give an observed (expected) bound of BF(H → inv.) < 0.40(0.40) at 95% C.L., and we show some of the implications for theories of dark matter which communicate through the Higgs portal.

Direct mass limits for chiral fourth-generation quarks in all mixing scenarios.

Present limits on chiral fourth-generation quark masses mb' and mt' are broadly generalized and strengthened by combining both t' and b' decays and considering a full range of t' and b' flavor-mixing scenarios with the lighter generations (to 1-‖V44‖2≈10(-13)). Various characteristic mass-splitting choices are considered. With mt'>mb' we find that CDF Collaboration limits on the b' mass vary by no more than 10%-20% with any choice of flavor mixing, while for the t' mass, we typically find stronger bounds, in some cases up to mt'>430 GeV. For mb'>mt', we find mb'>380-430 GeV, depending on the flavor mixing and the size of the mt'-mb' mass splitting.