# A fast and differentiable SMEFT likelihood

- Institution: CERN
- Department: Open meetings
- Speaker: Aleks Smolkovic
- Year: 2025
- Language: English
- Source: [CERN](https://videos.cern.ch/record/3021500)
- Licence: [CERN](http://copyright.web.cern.ch)
- Who can watch: Teens profiles and older (sources give no ratings; placed by Invisible Network from the source's own words)

The Standard Model Effective Field Theory (SMEFT) has proven to be a valuable framework for studying a broad class of models beyond the SM containing heavy degrees of freedom. The phenomenology of SMEFT can be developed systematically, incorporating essential effects such as Renormalization Group (RG) evolution and matching onto subsequent low-energy Weak Effective Field Theories (WET), enabling predictions for a wide range of observables. A likelihood function for the SMEFT that compares theory predictions to experimental data can be employed to identify patterns of deviation from SM predictions, but also to study the phenomenology of any specific model that can be matched onto SMEFT. In this talk, we present a new major version of the Python package smelli - a SMEFT likelihood, which incorporates a large number of observables, ranging from quark and lepton flavor physics, to Higgs physics, electroweak precision observables, beta decays, and high-mass Drell-Yan tails. It provides a fast and analytically differentiable global flavorful likelihood function for the SMEFT, WET, and for new physics models, with no assumptions on the flavor structure..

0:00:00 Slide 1

0:01:45 Slide 2

0:03:21 Slide 3

0:04:41 Slide 4

0:07:28 Slide 5

0:08:38 Slide 6

0:10:11 Slide 7

0:11:49 Slide 8

0:13:05 Slide 9

0:14:38 Slide 10

0:16:31 Slide 11

0:17:44 Slide 12

0:19:44 Slide 13

0:21:01 Slide 14

0:22:39 Slide 15

0:23:52 Slide 16

0:25:41 Slide 17

## Lectures

1. [A fast and differentiable SMEFT likelihood](https://invisiblenetwork.ai/learn/cern/3021500/a0a1c425e1f8), 29 min (CERN's player)