CERN Accelerating science

INSIGHTS: A new network of statistics for HEP and beyond

Following the success of the AMVA4NewPhysics network, a new 4-year Innovative Training Network funded by the Horizon 2020 program of the European Commission will develop new statistical methods and boost the use of machine learning (ML) to solve some of the most challenging problems in particle physics and society. The ITN, named INSIGHTS, is led by Glen Cowan, a professor at Royal Holloway University London, a member of the ATLAS collaboration, and a long-time expert of statistics techniques for data analysis, whose textbook on the topic is highly appreciated by the high-energy physics community. But speaking of books, one could not forget to mention that two other PIs of the network (Prof. Luca Lista, Univ. Naples, and Dr. Olaf Behnke, DESY) also authored successful textbooks in statistics for data analysis! That should clarify that the network has a really high training potential in this area.

The unprecedented amount of data at the Exa-Byte scale to be collected by the CERN experiments during the HL-LHC phase will require novel approaches to train and use ML models. Furthermore, the energy-frontier colliders currently in the process of being designed put even higher challenges in designing new statistics tools that will allow to disentangle and identify new physics in the foreseen high levels of background. Indeed, advanced statistical methods have proven to be key elements of recent advances in the field. INSIGHTS will enable significant further progress with particular emphasis on multivariate analysis, parametric modelling and Bayesian computation. Yet the development of modern state-of-the art ML methods and techniques will naturally lend itself to be also applied to problems in unrelated areas of research, of interest of climate science, internet data mining, and a number of applications in industry, following the tradition of high-energy physics

INSIGHTS follows the thread dug by AMVA4NewPhysics in pursuing and extending the above goals. It brings together a network of 11 beneficiaries (Royal Holloway London, Univ. Oslo, Max Planck Institute Munich, NIKHEF, INFN, KPMG, Univ. Naples, Univ. Lund, CERN, Pangea Formazione, Univ. Edimburgh) plus several partners from research centres, universities and the industry. 12 early-stage researchers (ESR) will join the ATLAS and CMS collaborations, digging deep in the datasets recently acquired by the experiments, or work on the development of ML tools for non-HEP applications. The INSIGHTS network has the principal aim of providing  them with advanced training in statistical methods and ML. Most of them will earn a Ph.D. in particle physics at the end of their 3-year involvement in the ITN; all will become highly-employable experts in data science.

More information on the INSIGHTS program is provided at the network web site, . From there, you may also get acquainted with the 12 ESRs and follow their blogging activities, which are just starting, following the steps of the 10 ESRs of AMVA4NewPhysics at