Colibri: a new tool for fast-flying PDF fits

5 May 2026, 11:30
10m
Sala IMPERIALE A, First Floor (Hotel Carlton)

Sala IMPERIALE A, First Floor

Hotel Carlton

Flash talk WG1 Structure Functions and Parton Densities WG1 Structure functions and parton densities

Speaker

Valentina Schutze (University of Cambridge)

Description

Despite its ability to quantify uncertainties probabilistically and naturally accommodate theoretical constraints, Bayesian inference has so far played a limited role in PDF fitting. In this talk I will present Colibri, an open-source code that provides a general and flexible tool for PDF fits. The code is built so that users can implement their own PDF model, and use built-in functionalities for a fast computation of observables. It grants easy access to experimental data, several error propagation methodologies, including the Hessian method, the Monte Carlo replica method, and an efficient numerical Bayesian sampling algorithm. To demonstrate the capabilities of Colibri, I will present a simple application: a polynomial PDF parametrisation. I further discuss how the functionalities illustrated in this example can be extended to more complex PDF parametrisations. In particular, Bayesian sampling in Colibri provides a framework for systematic model selection and model averaging, making it a valuable tool for benchmarking and combining different PDF parametrisations on solid statistical grounds.

Speaker confirmation Yes

Authors

Mark Costantini (University of Cambridge) Luca Mantani (Instituto de Fisica Corpuscular (IFIC), Universidad de Valencia-CSIC) James Moore (Lucy Cavendish College, University of Cambridge) Valentina Schutze (University of Cambridge) Maria Ubiali

Presentation materials