Using machine learning to explore new-physics models

Some results from our studies on using machine-learning to efficiently sample new-physics parameter space. Recall and Precision are ways to measure how well our model does at guessing if a new theory is excluded or not (this gets better the more data we include), while Entropy is a measure of how uncertain our model is of its predictions (the lower the better)

My latest paper preprint is out on the arXiv, submitted to the Sci Post Physics Core journal.

We used an “analysis recycling” method combined with some artificial intelligence to figure out if new physics theories we compatible with the existing data from the LHC, using about 5-10x less computing power than might previously have been needed. This means we can tackle much more complex models than previously possible.

I’ve been working on this paper for about a year with my former boss Jon Butterworth, and three talented ex-UCL undergraduates (Juan, Gustavs and Maria) who did most of the heavy lifting.

Watch this space as we apply this technique to probe ever more complex models!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: