Optimization examples

Loading...

Do curso por National Research University Higher School of Economics

Addressing Large Hadron Collider Challenges by Machine Learning

14 classificações

National Research University Higher School of Economics

Addressing Large Hadron Collider Challenges by Machine Learning

14 classificações

Curso 7 de 7 em Specialization Advanced Machine Learning

The Large Hadron Collider (LHC) is the largest data generation machine for the time being. It doesn’t produce the big data, the data is gigantic. Just one of the four experiments generates thousands gigabytes per second. The intensity of data flow is only going to be increased over the time. So the data processing techniques have to be quite sophisticated and unique. In this course we’ll introduce students into the main concepts of the Physics behind those data flow so the main puzzles of the Universe Physicists are seeking answers for will be much more transparent. Of course we will scrutinize the major stages of the data processing pipelines, and focus on the role of the Machine Learning techniques for such tasks as track pattern recognition, particle identification, online real-time processing (triggers) and search for very rare decays. The assignments of this course will give you opportunity to apply your skills in the search for the New Physics using advanced data analysis techniques. Upon the completion of the course you will understand both the principles of the Experimental Physics and Machine Learning much better.

Na lição

Detector optimization

This module covers several cases of detector design optimization in high energy physics experiments using Bayesian optimization with Gaussian processes.

Detector optimization7:05

Normal distribution3:37

Gaussian processes for regression4:26

Bayesian optimization. Part 14:36

Bayesian optimization. Part 25:09

Optimization examples3:29

Conheça os instrutores

Andrei Ustyuzhanin
Head of Laboratory for Methods of Big Data Analysis
HSE Faculty of Computer Science
Mikhail Hushchyn
Researcher at Laboratory for Methods of Big Data Analysis
HSE Faculty of Computer Science

In this video, we consider two examples of usage the

Bayesian optimization with Gaussian processes in real high energy physics experiments.

The first example is a muon shield optimization,

is the SHiP experiment at CERN.

The SHiP experiment is a new general purpose fixed target facility proposed at

the CERN SPS accelerator to search for new physics in

the largely explored domain of very weakly interacting particles.

The moun shield is a critical component of the SHiP experiment,

which deflects the high flux of muons produced in the target,

that would represent a very serious background for the particles searches,

away from the detector.

The shield consist of eight magnets and each magnet is parameterized by seven values.

The loss function depends on the physical performance of the shield and its weight.

So, it is a 42-dimensional optimization problem and

optimal moun shield geometry was found using 5000 iterations,

using the Bayesian optimization with Gaussian processes.

The optimal geometry of the shield is shown on the slide.

It's 25 percent lighter and as a result 25 percent cheaper.

Next example is about collisions simulation optimization.

Bayesian methods are used to optimize parameters of heavy-ion collision simulation.

Evaluating a simulation model for a single set of

parameters requires thousands of individual event simulations,

so direct optimization techniques,

like grid search, quickly become intractable.

This slide just shows simulated examples of

entropy density in the transverse plane for several,

typical ion-ion collision events.

With optimization, 300 initial points were generated in nine-dimensional parameter space.

These parameters are listed in the table on the slide.

For each of these 300 points,

about 10,000 of events were executed.

Figure on this slide demonstrates validation of Gaussian process emulator predictions.

Each panel shows predictions compared to explicit model calculations,

at a 50 validation design points.

This figure shows the quality of the simulation models with the optimal parameters.

Lines correspond to different simulation models with optimal parameters.

Points are data from the ALICE experiment.

In the bottom, it's a ratio of model calculations to data,

where the gray band indicates +/- 10 percent.

In this video, we considered that just several examples of using

Bayesian optimization methods in high energy physics.

In the programming assignment for this week,

you will be able to find optimal design for a simple tracking system.

Explore nosso catálogo

Registre-se gratuitamente e obtenha recomendações, atualizações e ofertas personalizadas.

O Coursera proporciona acesso universal à melhor educação do mundo fazendo parcerias com as melhores universidades e organizações para oferecer cursos on-line.

© 2018 Coursera Inc. Todos os direitos reservados.

Baixar na App Store

Baixar no Google Play