2 of 9

SFT in 2023: year of evolutionary changes

New GL, DGL
Evolution of group activities

Simulation: flattened structure (new PL: A Ribon)
Stacks (previously known as SPI), including Key4Hep (new PL: A Sailer)
Starting of SFTLabs: cradle for ideas, <20% member time

Currently: investigate possible role of Julia

Handle consequences of the NGT project

New ROOT PL: D Piparo

Revisitation of group communication channels

Consolidated public meeting structure
Re-established strategic role of Architect Forum
Direct representation in WLCG-MB and LHCC

NGT: Next Generation Triggers

3 of 9

SFT projects view

ALICE

ATLAS

CMS

LHCb

EP R&D

IT projects

HSF

FCC

…

Stakeholders

Simulation Tools

Geant4

Analysis And Processing Tools

ROOT

Software Distribution Tools

CernVM-FS

Software Stacks / SPI

Coordination Activities (HSF, EP R&D, …)

4 of 9

The one-day Plans-Of-Work and stakeholders feedback

Until 2023, the projects PoWs were presented in separate group meetings

From mid January to mid February

In 2023 we positively pioneered following-up the presentations with consolidated feedback from the LHC experiments
The one-day PoWs aims at

Facilitating the feedback collection/processing process
Having the feedback in time for the definition of each member objectives (MERIT)

The feedback will be discussed on the February 1st AF

The main results and corrections to the plans presented today will be presented in one of the following SFT regular meetings

5 of 9

Remarks about the one-day PoWs

Coordination activities not covered

They concern SFT participation to other activities w/ PoW discussed elsewhere

EP R&D

Starting Phase II, 2024-2028
SFT deeply involved in WP 7 activities
Projects will mention the connections/contributions in their presentations

6 of 9

Evolution in 2024 and focus on Machine Learning

Why?

The NGT discussions have underlined the ML relevance in the community
The KT Applied AI workshop also brought the attention to the need for our community to create a critical expertise across the lab and more structured approach
SFT is the natural place within EP where such efforts should converge and expertise build up

Machine Learning in SFT

Simulation (FastSim R&D) and ROOT (SOFIE, …) have ML components
Latest versions of ML tools available on lxplus/SWAN through the LCG stacks

7 of 9

ML¹: new horizontal activity in SFT

Expectations

Attract interest from developing programs, e.g. NGT
Act as synergising point for solutions developed in the community
Build-up critical expertise around available technologies�

Start from re-arranging current ML related activities from Simulation and ROOT�
Project Leader: Lorenzo Moneta

1: working project name

8 of 9

SFT new projects view