HSE and Yandex Hold Second Summer School in Machine Learning in High Energy Physics
On June 20-27, the HSE Faculty of Computer Science teamed up with the Yandex School of Data Analysis to carry out the Second International Summer School in Machine Learning in High Energy Physics. This year’s Summer School took place in Lund, Sweden, and saw the participation of 60 students and researchers from Austria, Great Britain, Germany, Italy, Russia, the U.S., Sweden, Switzerland, and more. Of these students, five came from the Higher School of Economics.
The Summer School’s programme offered two tracks – basic and advanced. The basic track focused on key topics needed to build classification models for data collected using modern-day particle detectors. Lecturers included Alexey Rogozhnikov, while practicums were held by Nikita Kazeev, Tatiana Likhomanenko, and Alexander Panin. The advanced track included topics such as particle trajectory restoration (Mikhail Guschin), methods for optimising real-time data selection algorithms (Tatiana Likhomanenko), and joint research technologies (Andrey Ustyuzhanin). Both tracks then turned into a discussion of how in-depth learning can be applied to physics experiments (lecture by Alexander Panin; practicums by Tatiana Likhomanenko and Alexey Rogozhnikov). To strengthen the data analysis methods covered at the lectures and seminars, two competitions were also organised on the Kaggle platform. The advanced tract worked on improving event filters, while the basic track worked on the search for decays of the Higgs boson. There were many physicists and machine-learning experts from all over the world invited to speak at the Summer School: Chase Shimmin (UC Irvine, U.S.), Gilles Louppe (NYU, U.S.), Vicens Gaitan (Grupo AIA R&D, Spain), Joaquin Vanschoren (TUE, the Netherlands), Michael Williams (MIT, U.S.). Two computer clusters were used at the school, the first of which was provided by Yandex (300 cores, 1.5 terabytes of memory), while the second came from a Finnish supercomputer centre (48 GPU cards).
Petr Zhizhin, First-year Applied Mathematics and Informatics Student
‘The trip to Sweden began when our student group got an invitation in the mail. I filed out the form, which asked about the different programming languages I knew, what I expected from my school, and why I should be selected for the event.
I’ve loved physics since I was in elementary school. It was one of the most difficult subjects available, and it always posed a challenge for me with all of the complex tasks it entails. Each year I participated in different physics Olympiads. I got a hundred on the Unified State Exam and even participated in the Faculty of Computer Science’s PhysTech Olympiad in physics.
So I simply had to apply for the School. It was an excellent opportunity for me to meet real physicists who work in the field of machine learning. When I got the call that I’d been accepted, I was beside myself with joy.
Classes at the School were taught by staff from the LAMBDA laboratory who work both within the faculty and at Yandex. It was a lot of work. I didn’t sleep more than five hours a single night of the Summer School. Over the course of a week, we looked at various algorithms such as linear regression, decision trees, random decision forests, boosting algorithms, and neural networks. After completing the School, I knew exactly which areas of mathematics I needed to study in the future.
For our track, the organisers selected a project dealing with machine learning in high-energy physics – the search for decays of the Higgs boson. To complete our task, we tried out a large number of different machine-learning algorithms. They didn’t all work out that well, but the main steps included random decision forests, XGBoost, XGBoost with Grid Search, and neural networks. I can definitely say that I was surprised when the neural networks yielded significantly better results than the other models.
After the Summer School, the instructors said that they were really impressed by our effort and that we were the youngest team at the school. They took note of the results we achieved in the competition and gave us commemorative prizes.’
In addition to Petr Zhizhin, a first-year student in the Applied Mathematics and Informatics programme, the Summer School also saw the participation of third-years Alexander Tiunov and Roman Schedrin of the Applied Mathematics and Informatics Programme, Ekaterina Rubtsova from the Software Engineering programme, and Artem Filatov from the Economics programme.
On December 1st and 2nd, the semifinals of the International Collegiate Programming Contest (ICPC) for teams from the Northern Eurasia region were held. HSE teams from St. Petersburg (8th place) and the Moscow-based Faculty of Computer Science (13th place) received first class certificates. Furthermore, the teams will represent HSE at the finals, which will be held in Portugal next year.
Exploring Political and Cultural Space of St Petersburg through the Summer School 'Topography of Imperial Power'
On a grey autumn day, it is always nice to warm up by reliving memories of summer adventures. This year, the balmy weather did not leave our city till mid-October, and a summer mood also lingered at HSE University – St Petersburg with the IV International Summer School 'The Topography of Imperial Power: Political and Cultural Space of Saint Petersburg' which ran from September 11 till October 2, 2018.
Professor Lev Shchur, Academic Supervisor of the master's programme in Supercomputer Simulations in Science and Engineering, and Professor Hai-Qing Lin, Director of Beijing Computational Science Research Center (CSRC) have signed a cooperation agreement for implementing a joint research project in computational physics and high performance computing.
In August 2018, the Faculty of Law at HSE Nizhny Novgorod hosted the annual summer school ‘Russian – Chinese Business: Cultural and Legal Aspects’. This year, the delegation from China was twice as big as last year, and the number of volunteers increased significantly, which clearly shows that summer schools with partner universities gain momentum!
The second Summer School on Deep Learning and Bayesian Methods (DeepBayes 2018) was held in Moscow from August 27 to September 1, 2018. The summer school was organized by HSE Centre of Deep Learning and Bayesian Methods and Samsung AI Center in Moscow. The lectures were taught by researchers from two centers-organizers, Skoltech, and Lomonosov Moscow State University.
The Russian Summer School, Masterchain, was held at the Higher School of Economics. It provided students with experience in working with this platform beyond the sphere of banking. The students collaborated on innovative projects for solving real business problems in the field of FinTech.
On June 9–15, the town of Pushkin near St. Petersburg, Russia hosted the Sixth International Summer School on Multidisciplinary Higher Education Research organized by the HSE Institute of Education (IOE). Bringing together a premier cohort of accomplished international academics and bright early-career scholars, the 2018 School aimed to explore new and deeper perspectives on how higher education can contribute to positive socio-economic change at all levels.
Every summer since 2007, the Center for Institutional Studies at the National Research University Higher School of Economics has been organizing the annual Russian Summer School on Institutional Analysis (RSSIA). Students and researchers from around the world convene in Moscow to enjoy a whole week of lectures and consultations, presentations, discussions and networking, as well as the opportunity to be inspired by the experts.
On September 12, Pham Cong Thang, who goes by Thang, arrived in Moscow to begin a post-doctoral fellowship at HSE’s International Laboratory of Deep Learning and Bayesian Methods (Faculty of Computer Science, Big Data and Information Retrieval School). Working under the supervision of Professor Dmitry Vetrov, Thang will focus primarily on text and image processing, and computer vision.
On July 17-23 the Third Machine Learning summer school organized by Yandex School of Data Analysis, Laboratory of Methods for Big Data Analysis at the National Research University Higher School of Economics and Imperial College London was held in Reading, UK. 60 students, doctoral students and researchers from 18 countries and 47 universities took part in the event.