Classifier surrogates: sharing AI-based searches with the world

In recent years, neural network-based classifica-
tion has been used to improve data analysis at collider exper-
iments. While this strategy proves to be hugely successful,
the underlying models are not commonly shared with the
public and rely on experiment-internal data as well as full
detector simulations. We show a concrete implementation of
a newly proposed strategy, so-called Classifier Surrogates, to
be trained inside the experiments, that only utilise publicly
accessible features and truth information. These surrogates
approximate the original classifier distribution, and can be
shared with the public. Subsequently, such a model can be
evaluated by sampling the classification output from high-
level information without requiring a sophisticated detector
simulation. Technically, we show that continuous normaliz-
ing flows are a suitable generative architecture that can be
efficiently trained to sample classification results using con-
ditional flow matching. We further demonstrate that these
models can be easily extended by Bayesian uncertainties
to indicate their degree of validity when confronted with
unknown inputs by the user. ... mehrFor a concrete example of tag-
ging jets from hadronically decaying top quarks, we demon-
strate the application of flows in combination with uncer-
tainty estimation through either inference of a mean-field
Gaussian weight posterior, or Monte Carlo sampling network
weights.