Defying 4D Gravity: Novel Searches for Signatures of String Theory Via Non-minimal Dark Sectors, Scalar Cascades, Noncommutative Black Holes, and Deep Learning

The existence of extra dimensions could lower the fundamental Planck scale to the low TeV scale and very excitingly allow string theory to be probed at the LHC. This thesis presents novel searches for signatures of string theory via three models that have never been searched for by any experiment: scalar cascades, noncommutative black holes, and non-minimal dark sectors. This thesis includes the first-time exploration of a novel final state. Deep learning is employed through the use of particle flow networks to learn jet substructure and design analysis regions. The background estimation is automated through the implementation of distance correlation in the neural network loss function. A novel jet reclustering algorithm is used to overcome the challenges associated with a soft final state.