Software R&D: Next Gen Simulation

Monday 27 Jun 2016, 17:30 → 18:30 Europe/Zurich

13/2-005 (CERN)

New simulation strategies are evolving next to full and "fast" simulation. This session will cover two examples. The discussion will cover how these future strategies could help with physicists' simulation needs.

17:30 → 18:00 FALCON: Towards an Ultra Fast Non-parametric Detector Simulator

Preliminary work towards a self-tuning non-parametric detector simulator that maps events at the generator level to reconstruction level.

Speaker: Dr Sergei Gleyzer (University of Florida (US))

Sergei_Gleyzer_Falcon_SFTForum.pdf

18:00 → 18:30 Learning to generate with adversarial networks

In this talk, we introduce generative adversarial networks (Goodfellow, 2014), a machine learning framework for building generative models through an adversarial learning process. Inspired from its recent success to model and generate natural images, we then explore and discuss how this framework could be used as a way to generate physics data.

Speaker: Gilles Louppe (New York University (US))

Gilles_Louppe_GANs_SFTForum.pdf