How does particle physics use deep neural networks to find particles? Does anyone use deep learning: RNN, CNN or any other architecture of deep neural networks to asses the standard model or to detect new or unseen particles? What's the status these days in this frontier? 
 A: The answer is no, for using it with the theoretical models, as the other answer says.
BUT
Neural networks are used as tools in deciding kinematic states in particle detectors, assigning momentum and energy to possible new particles and thus identifying  them  by using  a neural network "fit" even by the time I retired in 2000

A multivariate analysis based on the neural networks technique has been used to do the first inclusive measurement of the charmless semileptonic branching ratio of B-hadrons B → Xulvl in the ALEPH experiment at LEP.

In this link the use in high energy experiments is reviewed.

Artificial neural networks are the machine learning technique best known in the high energy physics community . Introduced in the field in 1988, followed by a decade of tests and applications received with reticence by the community , they became a common tool in high energy physics data analysis. Important physics results have been extracted using this method in the last decade.

A: The short answer is no. 
The verification process for the standard model involved comparing predictions furnished by that model with experimental data, mostly from particle accelerators, including some which were purpose-built to serve in this way. Neither the process of writing down the mathematical underpinnings for the model nor that of designing and building accelerators or the detectors used with them required neural networks. 
One of the frontiers in this field is the search for a candidate particle to furnish the missing mass needed to account for the dynamics of spiral galaxies and the clumping behavior of galaxy clusters, groups, and supergroups. It is not clear (to me anyway) how neural networks would be useful in this arena.
