From June 8 to June 12 we are organizing with the MODE Collaboration its fifth workshop, "V MODE Workshop on Differentiable Programming for Experiment Design". It will be held in Kolymbari, Greece, at the wonderful venue called "Orthodox Academy", a structure overlooking the Mediterranean sea in western Crete. Last September we had our fourth edition in Valencia, were about 80 participants (with an average age of less than 30 years!) gave super-interesting talks on recent development in experiment optimization and machine learning tools for fundamental research.
From Kolymbari I will fly to Cagliari, Italy for the second EUCAIF conference, which I am also co-organizing. EUCAIF is a coalition of physicists and astrophysicists who try to support the development of artificial intelligence tools for fundamental science research. Last year we were in Amsterdam, where over 280 attendees divided in several working groups to tackle many state-of-the-art topics in AI. I fully expect a super-interesting conference in Cagliari! As a convener of working group 2 (co-design) I will be busy with chairing and convening sessions there.
On June 20 I will fly from Cagliari to Monastir, in Tunisia, where I have been invited to give a lecture at a school in immunology that is organized with the collaboration of USERN, the network of 26,000 scientists who work for interdisciplinary science across borders for which I am currently serving as President. I will give a lecture there titled "".
From Monastir I will then travel to Cogne, where at an INFN school of particle detectors I will lecture on the topic of co-design of particle detectors with artificial intelligence. Cogne is a very pleasant mountain place, on the slopes of one of the highest peaks in the Alps, the Gran Paradiso. I hope I will be able to hike there during a free afternoon!
After Cogne, I am due in Louvain-la-Neuve as opposer of a Ph.D. defense, and then there follow some overdue vacations. I will resume my conference-going with the ICNFP conference, again in Kolymbari, at the end of July; and finally, I will be lecturing at the "Physics Days" conference organized in Lulea (Sweden) in mid August. Concerning that keynote talk, here is a quick synopsis:
In 2012 machine-learning tools achieved paradigm-shifting performance in image classification, but 2012 was also the year when the LHC collider at CERN discovered the Higgs boson. For the first time, a new fundamental particle was discovered with explicit use of machine learning tools. While until then machine learning was frowned upon as a valid tool by the physics community, in 2012 -almost overnight- it became a necessary instrument for data analysis in fundamental science. I call that event the first AI revolution in fundamental science.
Today we are ready for a second revolution, which will allow artificial intelligence to assist in the design of the complex instruments required to investigate matter at the shortest distance scales, by providing means for continuous scanning of the very high-dimensional space of design solutions of particle detectors. In order for that to happen, physicists have to team up with computer scientists to create the necessary interfaces (simulation tools, dimensionality reduction methods, optimization algorithms). Of particular importance is the concept of co-design, where hardware and software are optimized together, avoid misalignments that reduce the final performance of the resulting data collection and reduction pipelines. In this presentation I will describe the state of the art in these activities.
So, this seems already like a rather crowded plan for the summer. I hope I will be able to report more closely on each of these events as they unfold...
So, this seems already like a rather crowded plan for the summer. I hope I will be able to report more closely on each of these events as they unfold...
Comments