❌

Normal view

There are new articles available, click to refresh the page.
Before yesterdayIndico Feed [Home]

Delavnica: Uvod v ogrodje NVIDIA RAPIDS

✇Indico Feed [Home]
22 January 2025 at 08:00

Kratek opis: Udeleženci bodo spoznali odprtokodno ogrodje NVIDIA RAPIDS, ki vsebuje knjižnice za strojno pospešeno delo s podatki in strojno učenje. Predstavljena bo arhitektura ogrodja NVIDIA RAPIDS s poudarkom na knjižnicah cuDF, cuML in cuGraph. Udeleženci bodo na primerih preizkusili ogrodje NVIDIA RAPIDS in ga primerjali z rešitvami brez pospeševanja (knjižnici scikit-learn in NetworkX). V sklopu primerjave bodo na srednje veliki podatkovni zbirki naučili klasifikator in spoznali prednosti strojnega pospeševanja za potrebe strojnega učenja.

Podrobnejši opis: V tej delavnici bodo udeleženci spoznali odprtokodno ogrodje NVIDIA RAPIDS, ki omogoča strojno pospešeno obdelavo podatkov in izvajanje strojnega učenja ter napredne analize grafov, vse s ciljem hitrejšega izvajanja obdelav na velikih podatkovnih zbirkah. NVIDIA RAPIDS vključuje vrsto knjižnic, zasnovanih za izrabo zmogljivosti grafičnih procesorjev, ki omogočajo pospešeno analitiko in strojno učenje na obsežnih podatkih.

V sklopu delavnice bo podrobno predstavljena arhitektura ogrodja NVIDIA RAPIDS, pri čemer bo poudarek na treh ključnih knjižnicah: cuDF za hitro obdelavo podatkov, cuML za strojno učenje ter cuGraph za analizo grafov. Udeleženci bodo spoznali, kako te knjižnice omogočajo hitrejšo in učinkovitejšo obdelavo podatkov, v primerjavi s tradicionalnimi rešitvami, ki ne izkoriščajo pospeševanja, kot so scikit-learn za strojno učenje in NetworkX za analizo grafov.

Delavnica bo obsegala praktične primere, kjer bodo udeleženci primerjali izvajanje nalog strojnega učenja in analize grafov na srednje veliki podatkovni zbirki, pri čemer bodo uporabili NVIDIA RAPIDS in primerjali rezultate z rešitvami, ki temeljijo na CPE. Udeleženci se bodo naučili, kako hitro naučiti klasifikatorje na GPE ter raziskali prednosti strojnega pospeševanja pri obdelavi in analizi podatkov v primerjavi s tradicionalnimi pristopi. To bo udeležencem omogočilo boljše razumevanje prednosti uporabe ogrodja NVIDIA RAPIDS v realnih projektih strojnega učenja in podatkovne znanosti.

Zahtevnost: Napredna

Jezik: Slovenski

Termin: 22. 1. 2024 od 9.00 - 13.00

Omejitev števila udeležencev: 30

Virtualna lokacija: MS TEAMS

Priporočeno predznanje: Osnovno poznavanje programskega jezika Python, osnovno poznavanje knjižnic scikit-learn in NetworkX

Ciljna publika: Raziskovalci, inženirji, študenti, podatkovni znanstveniki, podatkovni analitiki 

Potek izobraževanja: Izobraževanje poteka na daljavo v okolju MS Teams. Udeleženci sodelujejo s pomočjo zvezkov Jupyter, ki jih odprejo na platformi Google Colab.

Na izobraževanju pridobljena znanja:

  • Razumevanje arhitekture ogrodja NVIDIA RAPIDS
  • Uporaba osnovnih in naprednih funkcij knjižnic cuDF, cuML in cuGraph
  • Obdelava podatkov na grafični kartici (cuDF)
  • Strojno pospeševanje algoritmov strojnega učenja (cuML in scikit-learn)
  • Strojno pospeševanje algoritmov za delo z grafi (cuGraph in NetworkX)
  • Uporaba ogrodja NVIDIA RAPIDS v praksi (priporočanje filmov in vektorsko iskanje)

 

Organizator:

Predavatelji:

Ime:Mladen Borovič
Opis:Mladen Borovič je asistent na Fakulteti za elektrotehniko, računalništvo in informatiko Univerze v Mariboru (UM FERI). Njegova raziskovalna področja so aplikacije umetne inteligence, priporočilni sistemi in iskalnike, sistemi za detekcijo podobnih vsebin, obdelava naravnega jezika in visokozmogljivo računalništvo.
E-mail:mladen.borovic@um.si 

 

AI@JSI Seminar - Josip Šarić - Label-efficient panoptic segmentation

✇Indico Feed [Home]
23 January 2025 at 12:00

Title: Label-efficient panoptic segmentation

Abstract: Panoptic segmentation provides a comprehensive understanding of visual scenes by assigning each pixel a semantic class label and, for objects, an instance ID. While highly effective, traditional methods rely heavily on large-scale annotated datasets, posing significant challenges for scalability and adaptability. Additionally, the process of annotating images with panoptic labels is both labor-intensive and time-consuming, highlighting the importance of label-efficient approaches. In this talk, I will present an overview of various label-efficient methods, focusing on my recent work in unsupervised domain adaptation and semi-supervised learning. I will also discuss the emerging field of open-vocabulary panoptic segmentation, which extends recognition capabilities to categories beyond the training taxonomy.

Lecturer: Josip Šarić, PhD

Short info on the lecturer: Josip Šarić is a postdoctoral researcher at the Faculty of Computer and Information Science, University of Ljubljana, supported by the SMASH postdoctoral program. Prior to this, he completed a PhD and two-year postdoc at the Faculty of Electrical Engineering and Computing, University of Zagreb. His research interests focus on computer vision and deep learning, with a particular emphasis on topics such as panoptic segmentation, open-vocabulary recognition, label-efficient learning, and related areas.

Information on AI@JSI seminars and previous recordings are available here: https://kt.ijs.si/aijsi-seminar/.

Delavnica: Visoko zmogljivi Python

✇Indico Feed [Home]
6 February 2025 at 09:00

Opis: Raziskovalci se pogosto soočajo z zahtevnimi računskimi izzivi, kot so analiza obsežnih podatkov, fizikalne simulacije, računska kemija, biologija, napovedovanje vremena, simulacije dinamike tekočin in podobno. Za reševanje teh izzivov se pogosto zatekajo k programskemu jeziku Python, ki ponuja bogato paleto knjižnic in orodij. Njegova slabost pa je lahko počasnost izvajanja, če ne uporabimo primernih pristopov k programiranju.

Na delavnici se bomo osredotočili na orodja in knjižnice v Pythonovem ekosistemu, s katerimi lahko izkoristimo računsko moč sodobnih večjedrnih procesorjev ter grafičnih pospeševalnikov v superračunalniškem okolju. Obravnavali bomo različne probleme, od operacij nad matrikami do učenja globokih nevronskih mrež, in poiskali ustrezne rešitve k pohitritvi izvajanja. 

Zahtevnost: Napredna

Jezik: Slovenski

Termin: 06. 02. 2025 od 10.00 - 16.00

Omejitev števila udeležencev: 30

Lokacija:

  • Fizična (Univerza v Ljubljani, Fakulteta za računalništvo in informatiko, Večna pot 113, Ljubljana, Predavalnica P19)
  • ZOOM 

 

Ciljna publika: raziskovalci, inženirji, študenti, vsi ki potrebujejo več računskih virov pri svojem delu

Zaželeno predhodno znanje:

  • opravljena delavnica Osnove superračunalništva,
  • razumevanje zgradbe računalniške gruče,
  •  delo preko odjemalca SSH (ukazna vrstica, prenašanje datotek),
  • osnovno poznavanje vmesne programske opreme Slurm,
  • osnovno znanje operacijskega sistema Linux in lupine Bash,
  • osnovno poznavanje programskega jezika Python.

 

Na izobraževanju pridobljena znanja:

  • Uporaba programskega jezika Python v superračunalniškem okolju
  • Uporaba knjižnic za vzporedno računanje v (RAPIDS, numba)
  • Uporaba knjižnice za porazdeljeno računanje DASK
  • Globoko učenje na superračunalniški gruči s knjižnico PyTorch

 

Organizator:

 

FRI logo

 

Predavatelji:

Ime: Davor Sluga
Opis: https://fri.uni-lj.si/sl/o-fakulteti/osebje/davor-sluga 
E-mail: davor.sluga@fri.uni-lj.si
Ime: Ratko Pilipović
Opis: https://www.fri.uni-lj.si/sl/o-fakulteti/osebje/ratko-pilipovic
E-mail: ratko.pilipovic@fri.uni-lj.si

 

International Masterclasses 2025

✇Indico Feed [Home]
19 March 2025 at 07:25

»International Masterclasses« iz fizike osnovnih delcev nudijo gimnazijcem enkratno priložnost, da se sami spoznajo s svetom kvarkov in leptonov, tako da izvedejo meritve na podatkih, zajetih pri eksperimentih v CERNu in drugih raziskovalnih centrih po svetu, da se srečajo z raziskovalci in se povežejo s svojimi vrstniki – dijaki iz drugih držav - in z njimi pregledajo rezultate ter izmenjajo mnenja. 

Dijaki bodo na enodnevnem dogodku s predstavitvami in delavnico, kjer bodo uporabili podatke, zajete z detektorjema ATLAS in  Belle II, spoznavali osnovne delce in sile, ki delujejo med njimi, ter uporabo tovrstnih delcev v sodobnih metodah zdravljenja s hadronsko radioterapijo. 

Celodnevni dogodek bo potekal na Institutu Jožef Stefan v Ljubljani. Dopoldne bomo raziskovalci z Instituta Jožef Stefan v Ljubljani, Fakultete za matematiko in fiziko Univerze v Ljubljani in Fakultete za kemijo in kemijsko tehnologijo Univerze v Mariboru predstavili fiziko osnovnih delcev, medicinsko fiziko in detektorje, ki jih uporabljamo pri naših raziskavah.

Na virtualnem sprehodu si bomo ogledali notranjost detektorja Belle II, vmes pa bo obilo priložnosti za pogovor z raziskovalci o tem, kako raziskujejo in kako preživljajo svoj čas v CERNu in na Japonskem. Izvedeli bomo lahko tudi nekaj več o tem, kakšne so povezave med raziskavami zgradbe vesolja in uporabo odkritij v medicinske namene. Ob koncu jutranjega dela delavnice bodo udeleženci prisluhnili še kratki predstavitvi enega od naših poslovnih partnerjev.

Kosilo za udeležence delavnice je predvideno v menzi Instituta Jožef Stefan. Za plačilo kosila (okoli 7 EUR) poskrbijo udeleženci sami. 

Po kosilu se bomo na popoldanski delavnici lotili iskanja neznanih kratkoživih delcev z uporabo podatkov, ki so bili zajeti z detektorjema ATLAS in  Belle II. Ogledali si bomo simulacijo radioterapije ter izbrali obsevalni načrt, ki bo učinkovito zavrl rast tumorja, pri tem pa obvaroval občutljive organe in obolelo tkivo. Po odkritjih se bomo skupaj z dijaki iz drugih raziskovalnih centrov po svetu povezali s kontrolno sobo eksperimentov v video konferenci.

Vabimo te, da se udeležiš poučnega in zabavnega dogodka. Dogodek bomo izpeljali v živo na Institutu Jožef Stefan.

Za udeležbo se registriraj preko zavihka na levi.

Mineral Detection of Neutrinos and Dark Matter (MDvDM) 2025

✇Indico Feed [Home]
19 May 2025 at 23:00

Mineral detectors have been proposed for a wide variety of applications, including searching for dark matter, measuring various fluxes of astrophysical neutrinos over gigayear timescales, monitoring nuclear reactors, and nuclear disarmament protocols; both as paleo-detectors using natural minerals that could have recorded the traces of nuclear recoils for timescales as long as a billion years and as detectors recording nuclear recoil events on laboratory timescales using natural or artificial minerals.

At this workshop, we will discuss the vast physics potential of mineral detectors, the progress in experimental studies, and the numerous challenges lying ahead on the path towards mineral detection. These include a better understanding of the formation and annealing of recoil defects in crystals; identifying the best classes of minerals and, for paleo-detectors, understanding their geology; modeling and control of the relevant backgrounds; developing, combining, and scaling up imaging and data analysis techniques; and many others.

List of Cofirmed Speakers

Alexey Elykov KIT Kohta Murase Penn State U.
Atsuhiro Umemoto KEK Lorenzo Caccianiga INFN Milan
Ayuki Kamada Warsaw U. Noriko Hasebe Kanazawa U.
Christopher Kelso U. of North Florida Patrick Huber Virginia Tech
Christian Wittweg U. of Zurich Patrick Stengel Jozef Stefan Institute
Daniel Ang U. of Maryland Shigenobu Hirose JAMSETC
Daniel Snowden-Ifft Occidental College Takenori Kato Nagoya U.
Emilie LaVoie-Ingram U. of Michigan Tatsuhiro Naka Toho U.
Igor Jovanovic U. of Michigan Vsevolod Ivanov Virginia Tech
Kai Sun U. of Michigan William McDonough Tohoku U.

 

During the last years, the MDvDM community has grown rapidly and gained attention. Small-scale experimental efforts focused on establishing various microscopic readout techniques are underway at institutions in Asia, Europe and North America. The third MDvDM workshop will bring together theoretical and experimental physicists, material scientists, and geologists to discuss the state of the art of the emerging field of Mineral Detection of Neutrinos and Dark Matter.

Previous MDvDM workshops:

  1. https://agenda.infn.it/event/32181/
  2. https://indico.phys.vt.edu/event/62/

Dan odprte znanosti 2024, EOSC tripartitni dogodek in Dan občanske znanosti

✇Indico Feed [Home]
3 December 2024 at 07:00

Dnevi odprte znanosti 2024 in EOSC tripartitni dogodek  – Odpiramo vrata prihodnosti znanosti z nacionalnim tripartitnim dogodkom EOSC!

Na Dnevnih odprte znanosti in Nacionalnem tripartitnem dogodku bomo gostili predstavnike Združenja EOSC, Evropske komisije, držav članic EU ter ključne domače in mednarodne strokovnjake s področja odprte znanosti, infrastrukture in vrednotenja znanosti. Potekali bodo od 3. do 5. decembra 2024 v  hotelu Four Points by Sheraton Ljubljana (Mons) v okviru Arnesove konference "Mreža znanja".

Prav tako bo v sklopu konference potekal tudi Dan občanske znanosti 2024. Poslanstvo odprte znanosti je odkrivanje novih znanstvenih spoznanj, hkrati pa tudi deljenje tega znanja med različnimi deležniki.

Na dogodke, ki bodo organizirani v sklopu Arnesove letne konference Mreža znanja, se je potrebno prijaviti. 

Spremljali boste lahko tudi prek video prenosov v živo.

 

 

 

Dnevi SLING

✇Indico Feed [Home]
3 December 2024 at 08:00

Vabimo vas na »Dneve slovenskega superračunalniškega omrežja« oz. »Dneve SLING« , ki bodo potekali v hotelu Four Points by Sheraton Ljubljana Mons. 

Spoznajte Slovensko superračunalniško omrežje (SLING) ter se seznanite z aktualnostmi s področja superračunalništva v Sloveniji in širše v Evropi.

Pester program vključuje predstavitve delovanja Nacionalnega kompetenčnega centra za superračunalništvo in centrov odličnosti na področju superračunalništva. Prav tako bomo predstavili primere učinkovite rabe zmogljivih superračunalnikov v industriji in akademski svetu ter razložili možnosti dostopa do teh sistemov za različne uporabnike.

Na dogodek, ki bo potekal v okviru projekta EuroCC 2 in Arnesove konference Mreža znanja, se je potrebno prijaviti. 

 

 Udeležba na dogodku in delavnicah je brezplačna. 

Poleg tega boste lahko "Dan slovenskega superračunalniškega omrežja" spremljali tudi prek prenosa v živo. 

Pridružite se nam in poglobite svoje razumevanje sveta superračunalništva.

Projekt EuroCC 2 financira Evropsko Skupno podjetje za evropsko visokozmogljivo računalništvo (JU) v okviru sporazuma o dodelitvi sredstev št. 101101903. JU podpirajo program Digitalna Evropa, Nemčija, Bolgarija, Avstrija, Hrvaška, Ciper, Češka republika, Danska, Estonija, Finska, Grčija, Madžarska, Irska, Italija, Litva, Latvija, Poljska, Portugalska, Romunija, Slovenija, Španija, Švedska, Francija, Nizozemska, Belgija, Luksemburg, Slovaška, Norveška, Turčija, Republika Severna Makedonija, Islandija, Črna gora in Srbija. Delovanje Nacionalnega kompetenčnega centra SLING sofinancira Ministrstvo za visoko šolstvo, znanost in inovacije.

Open Science Day 2024

✇Indico Feed [Home]
3 December 2024 at 08:00

 

Open Science Day 2024 and EOSC Tripartite Event

Following last year’s successful event, we invite you to join us again for the Open Science Day, taking place from the 3rd to 5th of December 2024.

The first day, December 3rd, will focus on international discussions and will be held in English. It will feature prominent guests, including representatives from the EOSC Association, the European Commission, EU member states, and experts in open science, infrastructure, and science evaluation. The event will run from 9:00 to 17:40.

The second day, December 4th, will focus on national topics and will be held in Slovene from 9:00 to 14:30.

The event will take place as part of the Mreža znanja conference at the Four Points by Sheraton Ljubljana (Mons), where the Citizen Science Day and Supercomputing Day will also be held.
 

 

 

SLING Days

✇Indico Feed [Home]
3 December 2024 at 08:00

We would like to invite you to the "Slovenian Supercomputing Network Days" or "SLING Days", which will take place at the  Four Points by Sheraton Ljubljana Mons Hotel.

Get to know the Slovenian Supercomputing Network (SLING) and learn about the latest developments in supercomputing in Slovenia and Europe.

The varied program includes presentations on the activities of the National Competence Center for Supercomputing and Centers of Excellence in Supercomputing. We will also present examples of effective use of powerful supercomputers in industry and academia, and explain the possibilities of access to these systems for different users.

Registration is required for this event, which is part of the EuroCC 2 project and the ARNES conference "Mreža znanja".

The event and the workshops are free of charge.

In addition, you will be able to follow the "Slovenian Supercomputing Network Day" via live stream.

Join us and deepen your understanding of the world of supercomputing.

 

EuroCC 2 project has received funding from the European High-Performance Computing Joint Undertaking (JU) under grant agreement No 101101903. The JU receives support from the Digital Europe Programme and Germany, Bulgaria, Austria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, Greece, Hungary, Ireland, Italy, Lithuania, Latvia, Poland, Portugal, Romania, Slovenia, Spain, Sweden, France, Netherlands, Belgium, Luxembourg, Slovakia, Norway, Türkiye, Republic of North Macedonia, Iceland, Montenegro, Serbia.
National Competence Centre SLING is co-funded by the Ministry of Higher Education, Science and Innovation, Republic of Slovenia.

 

Dan odprtih vrat slovenskih superračunalniških centrov ali Dan "exascale" računalništva

✇Indico Feed [Home]
16 October 2024 at 06:00

V oktobru bomo ponovno odprli vrata najzmogljivejših slovenskih superračunalnikov, vključno z EuroHPC Vega na Institutu informacijskih znanosti v Mariboru, ki se uvršča med najzmogljivejše superračunalnike na svetu.

Pridružite se nam in spoznali boste, kako superračunalniki vplivajo na naš vsakdan, od napovedi vremena do novih znanstvenih odkritij, ter se podrobno seznanili z njihovim delovanjem in uporabo. 

Za vas smo pripravili:

  • predstavitev tehnologije in razvoja superračunalništva v Sloveniji;
  • vodene oglede superračunalnikih sistemov.

 

18. oktobra 2024 bomo odprli vrata superračunalniških centrov širom Slovenije:

 

* Na IZUM-u bo Dan odprtih vrat EuroHPC Vege potekal v sredo, 16. 10., četrtek, 17.10., in petek, 18. 10. 2024.  
**Program podjetja Arctur je namenjen dijakom srednjih tehničnih šol in študentom.

Predstavitve bodo skupaj z ogledom potekale približno eno uro in bodo prilagojene različnim znanjem obiskovalcev. Pridružite se nam lahko tako posamezniki kot tudi organizirane skupine, kot npr. šolski razredi. Za vse udeležence je prijava obvezna.

Prijave in natančen razpored: TUKAJ 

Zaradi omejenih kapacitet bomo morali število obiskovalcev omejiti, zato pohitite s prijavami!

V sklopu dogodka bo potekala tudi spletna delavnica Generative AI with Diffusion Models v organizaciji Univerze v Mariboru. Obvezne prijave.

Oglejte si kratko video-predstavitev z obiska na IZUM-u ter posnetek spletne predstavitve primerov uporabe superračunalnikov.

Dan odprtih vrat sovpada z dnevom "exascale" superračunalništva, ki je 18. oktobra – dan, ko obeležujemo prihodnost superračunalnikov, z zmogljivostjo vsaj 1 exaFLOPS (10^18 operacij na sekundo).

 

 

Workshop: Generative AI with Diffusion Models

✇Indico Feed [Home]
18 October 2024 at 08:00

Description: At the workshop, we will present one of the most advanced areas of artificial intelligence - generative artificial intelligence. We will get to know all the components of generative diffusion models in more detail and in several successive steps we will create a test application for generating images based on text description.

The workshop is intended for both academic researchers and practitioners in the field of artificial intelligence who want to delve deeper into generative models and use them in their work. Throughout the day, we will explore concepts, conduct practical exercises and discuss the latest developments in this field.

After completing the workshop, participants will be able to obtain official certification from the NVIDIA Deep Learning Institute.

Detailed description: Thanks to the increasing capabilities of modern computer systems and advances in scientific theory, generative AI models have become more accessible than ever. Generative AI will play an important role in all branches of industry due to many applications such as creative creation of new content, rewarding existing data, simulations and planning, anomaly detection, new drug discovery, personalized recommendations and much more. In this workshop, we will delve into diffusion models for converting text to images.

The workshop will take place in the AWS cloud environment prepared by NVIDIA. Interactive access to powerful computing resources via a browser will be enabled. Each participant will have a powerful graphics accelerator at their disposal. The practical part of the workshop will be carried out using interactive learning documents, using the Python programming language and the PyTorch software library. The acquired knowledge can be easily transferred to other development environments.

After completing the workshop, participants will be able to take a short test and an additional programming task to earn official certification from the NVIDIA Deep Learning Institute. It will be possible to carry out the test and obtain the certificate at least six months after the end of the workshop.

Workflow: The workshop will take place in an interactive cloud environment with access via a browser.

Difficulty: Advanced

Language: English

Suggested prerequisites: Advanced knowledge of the Python programming language and programming libraries for building machine learning models. More detailed knowledge of the theoretical and practical aspects of deep neural networks.

Target audience: Students, academics and practitioners interested in generative models of artificial intelligence

Skills to be gained:

  • Hands-on experience in image generation from noise using U-Net neural network
  • Understanding the generative diffusion process based on denoising to improve the quality of the generated image
  • Understanding the differences between the diffusion process based on denoising and other generative models
  • Knowledge of procedures for controlling generated content by including context in the diffusion generative process


Max. number of participants: 30

Virtual location: MS Teams

Organisers:

 

 

 

Lecturers:

Name:Domen Verber
 

Domen Verber is an assistant professor at the Faculty of Electrical Engineering and Computer Science of the University of Maribor (UM FERI) and ambassador of the NVIDIA Deep Learning Institute for the University of Maribor and their HPC specialist. He has been dealing with HPC and artificial intelligence issues for more than 25 years.

 domen.verber@um.si, deep.learning@um.si

 

Name:Jani Dugonik
 

Jani Dugonik is an academic researcher at the Faculty of Electrical Engineering, Computer Science and Informatics of the University of Maribor (UM FERI). He has been working in the field of natural language processing and evolutionary algorithms for more than 10 years.

 jani.dugonik@um.si

Katarina Kokalj, "Structural analysis of partially fluorinated liquid ethanol"

✇Indico Feed [Home]
24 October 2024 at 12:00

Per- and polyfluoroalkyl substances (PFAS) are a group of synthetic fluorinated organic compounds that are indispensable in various applications, from consumer products to scientific research, due to their unique chemical and physical properties. Recently, these compounds have been found to be hazardous to both the environment and human health, and their remediation poses numerous challenges due to their stability and chemical inertness. Therefore, structural analysis of PFAS compounds is crucial to address this issue. Molecular dynamics (MD) simulations are an important tool in this regard as they allow the study of atomic behavior over time. In this work, we have focused on fluorinated alcohols. MD simulations were performed with the liquid 2,2,2-trifluoroethanol system (TFE) to determine how different force fields behave in describing the properties of fluorinated compounds. We have investigated six different force fields: TRAPPE, GROMOS-UA, GROMOS-AA, CHARMM, AMBER and OPLS. The suitability of these force fields for the prediction of structural and dynamic properties was evaluated by comparing the calculated results of simulation model systems with experimental X-ray scattering data, conformational analyses from the literature and certain thermodynamic quantities such as density and molecular diffusion coefficient. It was found that the TRAPPE, GROMOS-UA and GROMOS-AA force fields better describe the intermolecular correlations, while CHARMM, AMBER and OPLS are better suited to describe the intramolecular characteristics of TFE. The latter more accurately predicted the conformational forms of the molecules (gauche vs. trans), which we observed from the intramolecular spatial distribution functions and the average molecular end-to-end distances. On the other hand, the position of the maximum of the theoretical scattering curves was in better agreement with the experimental peak with the TRAPPE, GROMOS-UA and GROMOS-AA force fields.

Workshop: Advanced and non-textual tasks with ChatGPT

✇Indico Feed [Home]
4 November 2024 at 09:00

Date & time & place: November 4, 10.00-14.00, Zoom.

Description: In this workshop, attendees will learn how to expand the use of ChatGPT's textual interface to solve advanced tasks that are not usually associated with large language models.

ChatGPT is designed to be  a text-based assistant with several useful add-ons that are run externally to the GPT model. However, its textual interface alone can be used to solve much more demanding tasks, if we couple it with properly selected software packages or even (free) web tools. In this workshop attendees will learn how to use ChatGPT to:


- draw sketches and diagrams
- convert talks and lectures into study materials
- apply specific (yours or others) writing style to written materials
- generate presentation slides and posters from scientific papers
- generate graphic vector images from scans or screenshots
- generate graphic vector images from your speech
- generate graphical vector representation of alternative timelines in movies
- produce 3D rendered graphics
- bouncing ideas off ChatGPT, with sketches and code
(and perhaps others if time allows)

To fully participate in the workshop, the attendees are advised to register for free or paid ChatGPT account, to register for free account at https://www.overleaf.com, and to optionally download and install PovRay renderer, available for free at: https://www.povray.org/

Note: Large language models, such as those used in ChatGPT are notoriously unpredictable, even in absence of continuous upgrades, which can cause sudden rapid change in the models' performance. We cannot guarantee that attendees will be able to replicate the exact functionality that will be shown in the workshop. In case of problems with live ChatGPT interface, the tasks will be demonstrated using prompt logs from previous chats.

The duration of the workshop is set at 4 hours with the intention that we provide hands-on training for as many problems as possible and also answer your questions on how to approach your specific problem you may have!

Here are examples of output data we aim to generate, and the fifth example of fixing Excel table is linked under materials on the bottom of the page.

Difficulty: Intermediate

Language: English

Prerequisite knowledge: Basic familiarity with ChatGPT

Target audience: Students, (BSc, MSc, PhD), researchers, developers, engineers, general public 

Skills to be gained:

  • Gain the knowledge to use large language models for non-textual tasks
  • Gain the knowledge how to automate tedious textual tasks
  • Gain the knowledge on how to identify further opportunities to use ChatGPT in their own area of expertise

 

Workflow: The workshop is scheduled for a virtual session via Zoom. 

Virtual location: ZOOM - only registered participants will see ZOOM link!!

Organizer:

Lecturers:

Name:Janez Perš
 Janez Perš is an Associate Professor at the Faculty of Electrical Engineering at the University of Ljubljana. His research areas are computer, machine, and robotic vision, parallel and distributed systems, and human movement analysis. He teaches the courses on Embedded Systems and  Computer Vision.
E-mail:janez.pers@fe.uni-lj.si 

 

Name:Janez Križaj
 Janez Križaj is a researcher at the Faculty of Electrical Engineering at the University of Ljubljana. His research areas are deep learning, computer vision, biometrics, face recognition, pattern recognition, and image processing. 
E-mail:janez.krizaj@fe.uni-lj.si 

Emanuela Bianchi, "Modeling and simulating heterogeneously charged particles: from materials design to biological phenomena"

✇Indico Feed [Home]
5 November 2024 at 09:00

Globular proteins and recently synthesized colloids engineered with differently charged surface regions share a direction-dependent interaction characterized by a reduced bonding valence and a competition between like-charge attraction and opposite-charge repulsion. Understanding the large-scale behavior of heterogeneously charged particles is thus critical for exploring both biological processes, such as the liquid-liquid phase separation of globular proteins, and the assembly of target structures with specific properties at the nano- and micro-scale. 

Recently, we developed a simple coarse-grained model that, within well-defined limits, accurately reproduces the reference electrostatic potential derived from the linearized Poisson-Boltzmann framework [1]. This model is straightforward to implement in Monte Carlo and Molecular Dynamics (MD) simulations, enabling the exploration of how varying parameters -- such as net particle charge and surface charge pattern -- affect the self-assembly of these particles.

As illustrative examples, I will present our findings on how non-uniform electrostatics at the particle surface can influence the self-assembly of ordered phases [2] as well as the liquid-liquid phase separation [3].

[1] Soft Matter, 7, 8313 (2011), J.Chem. Phys., 142, 114905 (2015)
[2] Nanoscale, 9, 1956 (2017), Current Opinion in Colloid & Interface Science, 30, 18 (2017)
[3] arXiv:2401.10655, Soft Matter, 20, 7601(2024)

Delavnica: Osnove superračunalništva

✇Indico Feed [Home]
6 November 2024 at 08:00

Opis: Na delavnici se bomo seznanili z zgradbo računskih gruč in programsko opremo na njih ter zagnali svoje prve naloge. Naučili se boste razlikovati med prijavnimi vozlišči, računskimi vozlišči, ter sistemi za shranjevanje podatkov. Spoznali boste vlogo operacijskega sistema, vmesne programske opreme Slurm in uporabniških programov. Povezali se boste na prijavna vozlišča, prenašali datoteke na in iz superračunalnika, zaganjali naloge, s katerimi bomo obdelovali video posnetke, in spremljali izvajanje nalog.

Zahtevnost: Osnovna

Jezik: Slovenski

Termin: 06. 11. 2024 od 9.00 - 13.00

Omejitev števila udeležencev: 30

Virtualna lokacija: ZOOM 

Ciljna publika: raziskovalci, inženirji, študenti, vsi ki potrebujejo več računskih virov pri svojem delu

Na izobraževanju pridobljena znanja:

  • Razumevanje delovanja in zgradbe superračunalnikov
  • Uporaba vmesne programske opreme SLURM
  • Osnovna uporaba programskih okolij in vsebnikov
  • Upravljanje z datotekami in poganjanje nalog
  • Osnovna obdelava videoposnetkov

 

Organizator:

 

FRI logo

Predavatelji:

Ime:Davor Sluga
Opis:https://fri.uni-lj.si/sl/o-fakulteti/osebje/davor-sluga 
E-mail:davor.sluga@fri.uni-lj.si
Ime:Ratko Pilipović
Opis:https://www.fri.uni-lj.si/sl/o-fakulteti/osebje/ratko-pilipovic
E-mail:ratko.pilipovic@fri.uni-lj.si

 

Delavnica: Uporaba superračunalnika

✇Indico Feed [Home]
20 November 2024 at 15:15

Opis: Delavnica študente seznani z obsežnimi računalniškimi viri omrežja SLING, ki so ključni za napredne inženirske naloge. Udeleženci bodo pridobili dostop do superračunalniške gruče, se naučili osnovnih principov ter se seznanili s praktičnimi primeri vzporednega programiranja v jeziku C.

Podrobnejši opis: Sodobni računalniški algoritmi pogosto zahtevajo velike računalniške zmogljivosti. Pri nekaterih računalniških problemih, s katerimi se srečujejo inženirji, tudi zmogljiva delovna postaja hitro postane nezadostna ali pa ne omogoča hitrega izvajanja eksperimentov. Vendar pa imajo slovenski raziskovalci – vključno z doktorskimi študenti – dostop do več zmogljivih računalniških gruč (superračunalnikov), povezanih v slovensko superračunalniško omrežje SLING. Na tej delavnici boste (če želite) pridobili uporabniška imena na eni od računalniških gruč v omrežju SLING, spoznali osnovna načela dela s superračunalnikom in prevedli svoj prvi vzporedni program v programskem jeziku C.

Zahtevnost: Osnovna

Priporočeno predznanje: Ni zahtevano

Ciljna publika: Izobraževanje je namenjeno doktorskim študentom.

Na izobraževanju pridobljena znanja:

  • Arhitektura superračunalnikov, s poudarkom na računalnikih iz omrežja SLING;
  • Razlike v konceptu dela na delovni postaji (osebnem računalniku) in superračunalniku;
  • Nasveti za dobro izkoriščanje računalniških zmogljivosti superračunalnika;
  • Primer enostavne paralelizacije (množenje matrik, komunikacija med procesorji preko MPI).

 

Opis poteka izobraževanja: Delavnica bo potekala v fizični obliki na Fakulteti za elektrotehniko, Univerze v Ljubljani. Udeleženci bodo dostopali do superračunalniške gruče preko protokola SSH.

Lokacija: Fakulteta za elektrotehniko, Tržaška cesta 25, 1000 Ljubljana

Organizator:

Predavatelji:

Name:Janez Perš
 Janez Perš is an Associate Professor at the Faculty of Electrical Engineering at the University of Ljubljana. His research areas are computer, machine, and robotic vision, parallel and distributed systems, and human movement analysis. He teaches the courses on Embedded Systems and  Computer Vision.
E-mail:janez.pers@fe.uni-lj.si 

 

Name:Janez Križaj
 Janez Križaj is a researcher at the Faculty of Electrical Engineering at the University of Ljubljana. His research areas are deep learning, computer vision, biometrics, face recognition, pattern recognition, and image processing. 
E-mail:janez.krizaj@fe.uni-lj.si 

Abhishek Mohapatra: One Born Oppenheimer Effective Theory to rule all Exotics

✇Indico Feed [Home]
21 November 2024 at 14:30

The XYZ exotic states discovered in the hadronic sector with two heavy quarks constitute one of the most important open problems in particle theory. In this talk, I show that an effective field theory derived from QCD, the Born Oppenheimer effective field theory (BOEFT), can describe exotics of any composition. I show the results of general Schr\"odinger coupled equations for arbitrary angular momentum of the light degrees of freedom. The coupled equations describe hybrids, tetraquarks, pentaquarks, doubly heavy baryons, and quarkonia in leading order, including nonadiabatic terms. Additionally, I also present the results of the predicted multiplets, corresponding selection rules, and expressions of the nonperturbative gauge invariant correlators, which are the input of the BOEFT: static energies, generalized Wilson loops, gluelumps, and adjoint mesons that should be calculated on the lattice. Moreover, I show for the first time new results on the behavior of static energies at short distances and mixing with the threshold at long distances based on BOEFT. As an application of this BOEFT, I show results for the hybrid spectrum, hybrid decays to quarkonium and some preliminary results for X(3872) and T_cc (3875).

Workshop: Python binding for C/C++

✇Indico Feed [Home]
28 November 2024 at 12:00

Description: Scientific and math software is written in C/C++ which often requires longer development time when combining a larger set of such libraries. To better utilize such libraries it is often convenient to interface them to a modern interpreted language such as Python.

We will learn how to bind C/C++ code in Python 3, starting by creating a Python object that will utilize one of the standard C/C++ containers, the vector. Continuing we will then take an existing C/C++ library and prepare it to be called from Python. Finally we will demonstrate a workflow where two different libraries are combined in a single Python 3 script.

Difficulty: Basic

Language: English

Prerequisite knowledge: Basic knowledge of Linux, the Terminal, Python and C/C++

Target audience: The workshop is intended for beginners and more experienced people interested in utilizing C/C++ code from Python

Skills to be gained:

  • Identifying useful libraries, and how to combine them in a workflow
  • Basics of libraries for Python
  • Binding C/C++ libraries in Python

 

Workflow: The training is online, in the afternoon. The workshop will combine lecture and practical parts, where your own laptop suffices (however, if need users can also make use of the NSC Cluster at the Jožef Stefan Institute (as an educational member of SLING)).

Virtual location: ZOOM (only registered participants will see ZOOM link)

Organizer:

Univerza v Ljubljani v leto 2024 ...

Lecturer:

Name:Luka Leskovec
Description:Scientist and educationalist involved in theoretical physics and supercomputing
E-mail:luka.leskovec@fmf.uni-lj.si

Workshop: Introduction to High-Throughput-Computing: SnakeMake

✇Indico Feed [Home]
12 December 2024 at 12:00

Excerpt: We will learn the basics of SnakeMake, how to translate a basic workflow, which is partly trivially parallelizable and in parts not, into SnakeMake, and how to utilize HPC resources in conjunction with SnakeMake. We will take a look at a generic workflow, consisting of a set of tasks, with varying interdependencies; we will learn how to identify these interdependencies between our tasks and how to design a workflow diagram around them. 

Description: We will learn the basics of SnakeMake, which is an open-source tool that allows users to describe complex workflows with a hybrid of Python and shell scripting. By transferring our workflow diagram into a SnakeMake we will learn how to automate our workflow. Initial development of our workflow in SnakeMake will be in a terminal setup, as we do generically in our day-to-day lives, however we will then learn how to scale up and perform these tasks on a local supercomputing cluster.

Difficulty: Basic

Language: English

Prerequisite knowledge: Basic knowledge of Linux, the Terminal and Python

Target audience: The workshop is intended for beginners and more experienced people interested in optimizing and automating their workflows and ultimately transferring them to High-Performance Computing machines.

Skills to be gained:

  • Workflow design, identification of interdependencies
  • Basics of SnakeMake
  • Transfering a generic workflow into SnakeMake
  • Testing our SnakeMake in an interactive environment, and upscaling it for use on a supercomputing cluster

 

Workflow: The training is live, in the afternoon. Part of the training will be in the form of a lecture and part will be a practical session, where we will make use of the NSC Cluster at the Jožef Stefan Institute (as an educational member of SLING).

Virtual location: ZOOM (only registered participants will see ZOOM link)

Organizer:

Univerza v Ljubljani v leto 2024 ...

Lecturer:

Name:Luka Leskovec
Description:Scientist and educationalist involved in theoretical physics and supercomputing
E-mail:luka.leskovec@fmf.uni-lj.si

Delavnica: Vsebniki na superračunalnikih

✇Indico Feed [Home]
18 December 2024 at 08:00

Opis: Raziskovalci se pogosto spopadajo z velikimi računskimi izzivi, na primer pri analizi velikih podatkov, fizikalnih simulacijah, računski kemiji, računski biologiji, napovedovanju vremena, simulacijah dinamike tekočin ipd. Za reševanje mnogih problemov je pogosto na voljo ustrezna programska oprema, ki pa jo je potrebno prilagoditi za izvajanje na izbranem superračunalniku.

Na delavnici si bomo ogledali več načinov nalaganja programske opreme: v domačo mapo, preko okoljskih modulov in vsebnikov. Spoznali se bomo s konceptom virtualnih strojev in vsebnikov ter osvetlili razlike med zasnovo vsebnikov Docker in Apptainer. Naučili se bomo uporabiti že pripravljene vsebnike in na praktičnih primerih spoznali, kako zgraditi enostaven vsebnik Apptainer ter ga zagnati v superračunalniškem okolju. V nadaljevanju si bomo ogledali, kako v vsebnik vključiti podporo za grafične pospeševalnike in procesiranje na več vozliščih.

Delavnica bo praktično usmerjena, vaje bomo izvajali na modernem sistemu HPC.

Zahtevnost: Napredna

Jezik: Slovenski

Termin: 18. 12. 2024 od 9.00 - 13.00

Omejitev števila udeležencev: 30

Virtualna lokacija: ZOOM

Priporočeno predznanje: 

  • opravljena delavnica Osnove superračunalništva,
  • razumevanje zgradbe računalniške gruče,
  • delo preko odjemalca SSH (ukazna vrstica, prenašanje datotek),
  • osnovno poznavanje vmesne programske opreme Slurm,
  • osnovno znanje operacijskega sistema Linux in lupine Bash
  • osnovno poznavanje programskega jezika Python

Ciljna publika: raziskovalci, inženirji, študenti, vsi ki potrebujejo več računskih virov pri svojem delu

Na izobraževanju pridobljena znanja:

  • poznavanje vmesne programske opreme Slurm
  • razumevanje okoljskih modulov in vsebnikov
  • uporaba obstoječih vsebnikov Docker in Apptainer
  • gradnja lastnih vsebnikov Apptainer za izvajanje izbranih programov na superračunalniški gruči
  • raba različnih računskih virov v okoljskih modulih in vsebnikih (procesorska jedra, grafični pospeševalniki, vozlišča)

 

Organizator:

FRI logo

Predavatelji:

Ime:Davor Sluga
Opis:https://fri.uni-lj.si/sl/o-fakulteti/osebje/davor-sluga 
E-mail:davor.sluga@fri.uni-lj.si
Ime:Ratko Pilipović
Opis:https://www.fri.uni-lj.si/sl/o-fakulteti/osebje/ratko-pilipovic
E-mail:ratko.pilipovic@fri.uni-lj.si

 

Chandan Hati: Nonstandard Nucleon Decays

✇Indico Feed [Home]
28 June 2024 at 12:00

Nucleon decays are generic predictions of many well motivated theories, including those based on the unification of forces and supersymmetry. I will discuss how nonstandard nucleon decays offer a unique opportunity to broadly probe light new particles beyond the Standard Model with masses below few GeV, including axion-like particles, dark photons, sterile neutrinos, and scalar dark matter. Conventional searches can misinterpret and even completely miss such new physics. I will discuss a general strategy based on momenta of visible decay final states to probe these processes, offering a rich physics program for existing and upcoming experiments such as Super-Kamiokande, Hyper-Kamiokande, DUNE, and JUNO.

Mihaela Marinescu, "Measurement of the W and Z cross section at 13.6 TeV with ATLAS"

✇Indico Feed [Home]
4 July 2024 at 12:00
Measurements the W and Z boson cross section provide important tests of the Standard Model, offering insights into the underlying dynamics of strongly interacting particles and proton structure. The large cross sections of these processes, together with clean experimental signatures achieved through their leptonic decays, allow for excellent experimental precision reaching the percent level, and in the case of cross-section ratios, the sub-percent level.  This talk presents the recent measurement of the W and Z boson cross sections, their ratios and the ratio of the ttbar and W boson fiducial cross sections, using proton-proton collision data obtained at the new centre-of-mass energy of 13.6 TeV by the ATLAS experiment. The data used in this measurement correspond to an integrated luminosity of 29 inverse fb and was collected in 2022, at the beginning of the Run-3 data-taking period. Finally, results are compared with state-of-the-art Standard Model theoretical predictions calculated using several different PDF sets.

Shoji, Baratella: The Standard Model lifetime is slightly shorter

✇Indico Feed [Home]
25 June 2024 at 12:00

We revisit the decay rate of the electroweak vacuum in the Standard Model with the full one-loop prefactor. We focus on the gauge degrees of freedom and derive the degeneracy factors appearing in the functional determinant using group theoretical arguments. Our treatment shows that the transverse modes were previously overcounted, so we revise the calculation of that part of the prefactor. The new result modifies the gauge fields' contribution by 7%7% and slightly decreases the previously predicted lifetime of the electroweak vacuum, which remains much longer than the age of the universe. Our discussion of the transverse mode degeneracy applies to any calculation of functional determinants involving gauge fields in four dimensions.

Seminar: My life with the TACC scheduler

✇Indico Feed [Home]
1 July 2024 at 12:00

Description: In high-performance computing (HPC), efficient job scheduling is crucial for maximizing resource utilization and minimizing wait times. The Texas Advanced Computing Center (TACC) has developed a sophisticated scheduler designed to enhance the performance of HPC. After a brief overview of the logic behind the TACC scheduler, this talk will delve into a users' view and use of this very powerful tool.

Difficulty: Basic

Language: English

Target audience: The workshop is intended for anyone wanting to gain practical knowledge of HPC tools.

Maximum number of participants: no limit

Location: online via Zoom

Organizer:

Univerza v Ljubljani v leto 2024 ...

Lecturer:

Name:Luka Leskovec
Description:Scientist and educationalist involved in theoretical physics and supercomputing
E-mail:luka.leskovec@fmf.uni-lj.si

Seminar: Computer Architecture for Machine Learning

✇Indico Feed [Home]
5 June 2024 at 10:00

Deep learning and large language models have become one of the most promising research areas in artificial intelligence. However, the computational requirements for implementing these methods are also growing exponentially. In this lecture, we will discuss current trends in computer system architectures designed for machine learning, focusing on the challenges and opportunities this field faces. We will present some popular computer architectures that are used for efficiently executing machine learning algorithms.

Jezik: Angleščina

Termin: 5. 06. 2024 od 12.00 do 13.00

Lokacija: 

  • Naslov: Fakulteta za računalništvo in informatiko, Večna pot 113, Ljubljana
  • Prostor: Predavalnica P03

 

Organizator:

 

FRI logo

Predavatelji:

Ime:Dr. Nader Bagherzadeh
Opis: 
E-mail:nader@uci.edu

 

 

 

Seminar: Superračunalniki obožujejo poglobljeno učenje

✇Indico Feed [Home]
30 May 2024 at 09:00

Opis: Superračunalniki že desetletja igrajo pomembno vlogo pri napredku v znanosti in tehnologiji. Načrtovanje strojne in programske opreme zanje je vedno predstavljalo velik izziv za računalniške strokovnjake. Z napredkom strojnega učenja, predvsem z naraščanjem potreb po učenju globokih modelov nevronskih mrež, pa so računalniški strokovnjaki postali tudi njihovi veliki uporabniki. V predavanju bomo skozi razvoj superračunalništva v Sloveniji in širše pogledali, kako je prišlo do zasuka, ki je prinesel tudi spremembe v načrtovanju sistemov. Pri tem se bomo osredotočili na vlogo grafičnih pospeševalnikov, njihovo zasnovo in vedno nove prilagoditve, od dodajanja posebnih računskih enot za matrično množenje do uvajanja novih predstavitev števil.

Jezik: Slovenski

Termin: 30. 05. 2024 od 11.00 do 12.00

Omejitev števila udeležencev: brez

Lokacija: 

  • Naslov: Fakulteta za računalništvo in informatiko, Večna pot 113, Ljubljana
  • Prostor: Predavalnica P03

 

Organizator:

 

FRI logo

Predavatelji:

Ime:Uroš Lotrič
Opis:https://fri.uni-lj.si/sl/o-fakulteti/osebje/uros-lotric
E-mail:uros.lotric@fri.uni-lj.si

 

 

 

Delavnica: CFD on HPC – OpenFOAM primer

✇Indico Feed [Home]
27 May 2024 at 07:00

The three-day workshop is aimed at researchers, engineers, students and others interested in CFD and OpenFOAM (OF). Basic and advanced applications of the open source fluid simulation system OpenFOAM will be presented. The workshop will take place over three consecutive mornings via the Zoom system (zoom link will be published shortly).

Description: The three-day workshop includes a presentation of the basic and advanced OpenFOAM usage, with explanation of programming of new sets within the framework. The basic part includes steps of making a mesh, setting boundary / initial conditions, running the problem on the HPC system and reviewing the results. The advanced part covers more complex cases with advanced application usage. What is not covered in the OpenFOAM can be upgraded within the software environment with the implementation of your own code. Showcases and hands-on cases will be presented.

For students, academics and industry professionals who want to learn about the OpenFOAM open source CFD platform.

The training takes place online, in the morning. The interactive work is done through remote access to the HPC system at ULFS.

After the training you will : 

  • Be able to connect to HPC@ULFS with NoMachine client and work in HPC Linux environment,
  • Understand the theoretical background of the Computational Fluid Mechanics (CFD), esspecially of the Finite Volume Method (FVM),
  • Be able to set up CFD mesh using different open source programs for CFD mesh design (OF – Block Mesh, GMSH),
  • Be able to setup complete OF case (mesh, pysical model, inital and boundary conditions, ...),
  • Be able to setup and run various OF cases in parallel on an HPC cluster,

Be able to preview and post-process OF results.

Difficulty: Intermediate 

Language: According to the applications

Tridnevna delavnica je namenjena raziskovalcem, inženirjem, študentom in drugim, ki jih zanima tematika CFD z uporabo odprtega programskega CFD okolja OpenFOAM (OF). Predstavljena bo osnovna in napredna uporaba, s prikazom na primerih v treh zaporednih dopoldnevih, preko Zoom Sistema.

Opis: Tridnevni sklop obsega predstavitev osnovnega dela uporabe, naprednega dela uporabe in programiranja novih sklopov znotraj OpenFOAM. V osnovnem delu je predstavljen osnovni del uporabe, kjer so predstavljeni koraki izdelave mreže, robnih/začetnih pogojev, poganjanja problema na HPC sistemu in pregledovanja rezultatov. Napredni del pokriva bolj kompleksne primere z dodanimi naprednimi možnostmi uporabe. Kar v sistemu ni pokrito se da nadgraditi v programskem okolju. Prikazana bo tudi uporaba programskega dela na osnovnih primerih.

Izobraževanje je namenjeno študentom, zaposlenim v akademiji in gospodarstvu, ki si želijo spoznati odprto kodno CFD platformo OpenFOAM.

Izobraževanje poteka preko spleta, v dopoldanskem času. Interaktivno delo poteka preko oddaljenega dostopa do HPC sistema na ULFS. Zoom povezava bo objavljena v kratkem.

Zahtevnost: napredna

Priporočeno predznanje:
Osnove Linux operacijskega sistema in osnove mehanike tekočin ter osnove Python programiranja.

Jezik: glede na prijave

 

Organizer/Organizator

LECAD laboratorij, Fakulteta za strojništvo, Univerza v Ljubljani

 

Predavatelji

Ime:Dr. Pavel Tomšič
Opis:He is a research assistant at ULFME and is well qualified for several HPC related topics. He is actively involved in efforts to raise competencies in the field of supercomputing, such as the Partnership for Advanced Computing in Europe (PRACE). He is also coordinator of Erasmus + project SCtrain - a strategic partnership for the transfer of knowledge from supercomputing between Slovenia, Austria, the Czech Republic and Italy. As part of the EuroHPC project for the establishment of European National Competence Centers in the field of supercomputing (EuroCC), he is the champion for Training and Skills Development for NCC Slovenia.
E-naslov:pavel.tomsic@fs.uni-lj.si
Ime:Dr. Aleksander Grm
Opis:Aleksander Grm graduated with a Bachelor's degree in Physics from the Faculty of Mathematics and Physics at the University of Ljubljana. He then completed a Master's degree in Applied Mathematics at ICTP/SISA in Trieste, Italy. After the MSc, he continued his studies at the University of Kaiserslautern in Germany and obtained a PhD in Industrial Mathematics. After the PhD, he worked partly in academia and fully in industry. In 2014, he moved to the University of Ljubljana to work in basic and applied research and to teach young people mechanics and mathematics at the engineering level.
E-naslov:aleksander.grm@fs.uni-lj.si 

Innes Bigaran: Leveraging intermediate resonances to probe CP-violation at colliders

✇Indico Feed [Home]
28 May 2024 at 12:00

We study the phenomenological implications of interference between tree-level contributions to decays to three-body final states. CP-violation in these decays arise via different virtualities of intermediate resonances, and the presence of SM and/or new CP phases in a model Lagrangian. We highlight the manifestation of this CP-violating observable in the interference between SM charged-current decays and new-physics contributions to effective vector interactions in the EFT and in a UV-complete model featuring a scalar leptoquark, $S_1$. We present a novel method for searching for CP-violating new physics effects utilising the differential cross sections of $pp\to d_i \ell\nu_\ell$ decays, with a detailed study of $pp\to b \tau\nu_\tau$ simulating signal and background.

Workshop: Parallel R

✇Indico Feed [Home]
29 May 2024 at 11:00

Description: This training is focused on different approaches to how the R code can be parallelised, going from local and single node approach with parallel package to multi-node parallelisation with Rmpi package.

The second day is focused on increasing computation efficiency by introducing Rcpp for seamless integration of C++ code into R code. A simple example of CUDA usage with Rcpp will be shown.

Difficulty: Intermediate

Language: English

Prerequisite knowledge: Basic knowledge of programming in R.

Basic knowledge of working with Linux terminal is helpful.

The second day will introduce a combination of C++ and CUDA code with R. No prior knowledge is really needed as the examples will be kept simple and explained.

Skills to be gained:

  • Local or single-node parallelisation of R code with parallel package
  • Multi-node parallelisation of R code with Rmpi package
  • How to speed up code using C++ with Rcpp
  • How to leverage CUDA parallelisation on GPU with Rcpp

 

Workflow: We will use a remote instance of the RStudio server, so only an internet connection and web browser are needed.

Location: Online (ZOOM) - Only registered participants will receive the Zoom link.

Lecturers:

Janez Povh is a full professor of applied mathematics at the University of Ljubljana, Slovenia, and since 2022, he has been the managing director of Rudolfovo Institute in Novo Mesto, Slovenia. His research focuses on developing state-of-the-art algorithms to solve hard computational problems, mostly from combinatorial optimisation and data science. He is mostly interested in computing exact solutions (global optima) but adapts his results to get local optima. He analyses the algorithms theoretically and implements them into high-performance code that can harness the capabilities of the best available supercomputers.

Tomáš Martinovič obtained his Ph.D. in computational sciences at IT4Innovations, VSB - Technical University of Ostrava in 2018. From 2015 to 2018, he worked in a team focused on analysing complex dynamic systems, where he worked on scalable implementations of algorithms from the field of nonlinear time series analysis. Since the start of 2022, he has led a team focused on machine learning/AI and operations research with the defined objective of research and knowledge transfer in cooperation with industry.

Mario Reig Lopez: Testing the far UV of the Standard Model with axions

✇Indico Feed [Home]
30 May 2024 at 12:00
In this talk I will discuss axion couplings to gauge bosons in 4-dimensional Grand Unified Theories and in (heterotic) String Theory. The topological nature of these couplings allows them to be matched from the UV to the IR, and the ratio of the anomaly with photons and gluons for any axion is fixed by unification.  I will show that this implies that there is a single axion — the QCD axion — with an anomalous coupling to photons. Other light axion-like particles can couple to photons by mixing with the QCD axion, and necessarily lie to the right of the QCD line prediction in the mass-coupling plane. As a result, a discovery of an axion to the left of the QCD line can rule out simple GUTs and some heterotic string compactifications.
Based on: 2206.07053 and future work.

CERN 70 - Public Lecture: From Subatomic to Superheroes: Particle Physics and Possibilities / Od subatomskega do superjunakov: fizika delcev in njene priložnosti

✇Indico Feed [Home]
31 May 2024 at 17:00

Prijave in več informacij o dogodku na https://www-f9.ijs.si/sl/novice-dogodki/cern70-od-subatomskega-do-superjunakov/.

Govorili bomo, kako je stremenje k razumevanju najmanjših gradnikov vesolja pripeljalo do prelomnih odkritij in tehnološkega napredka. Spomnili se bomo nekaj pomembnih mejnikov v delovanju Evropskega laboratorija za fiziko delcev CERN, od rojstva svetovnega spleta do odkritja Higgsovega bozona. Predstavili bomo najsodobnejše eksperimente, ki oblikujejo naše razumevanje vesolja, in kako ti vplivajo na širšo družbo.

We’ll explore how the quest to understand the smallest components of the universe has led to groundbreaking discoveries and technological advancements. This talk will recall some of CERN's significant milestones, from the birth of the World Wide Web to the discovery of the Higgs boson, and discuss the cutting-edge experiments that have help shape our understanding of the cosmos and how they can have wider impact on society.

Mike Lamont, Director for Accelerators and Technology (Biography)

Delavnica: Visoko zmogljivi Python

✇Indico Feed [Home]
7 June 2024 at 07:00

Opis: Raziskovalci se pogosto soočajo z zahtevnimi računskimi izzivi, kot so analiza obsežnih podatkov, fizikalne simulacije, računska kemija, biologija, napovedovanje vremena, simulacije dinamike tekočin in podobno. Za reševanje teh izzivov se pogosto zatekajo k programskemu jeziku Python, ki ponuja bogato paleto knjižnic in orodij. Njegova slabost pa je lahko počasnost izvajanja, če ne uporabimo primernih pristopov k programiranju.

Na delavnici se bomo osredotočili na orodja in knjižnice v Pythonovem ekosistemu, s katerimi lahko izkoristimo računsko moč sodobnih večjedrnih procesorjev ter grafičnih pospeševalnikov v superračunalniškem okolju. Obravnavali bomo različne probleme, od operacij nad matrikami do učenja globokih nevronskih mrež, in poiskali ustrezne rešitve k pohitritvi izvajanja.

Zahtevnost: Napredna

Jezik: Slovenski

Termin: 07. 06. 2024 od 9.00 - 14.00

Omejitev števila udeležencev: 30

Lokacija: 

  • Naslov: Fakulteta za računalništvo in informatiko, Večna pot 113, Ljubljana
  • Prostor: Predavalnica P04

 

Virtualna lokacija: https://uni-lj-si.zoom.us/j/92053788501 

Priporočeno predznanje: 

  • opravljena delavnica Osnove superračunalništva,
  • razumevanje zgradbe računalniške gruče,
  • delo preko odjemalca SSH (ukazna vrstica, prenašanje datotek),
  • osnovno poznavanje vmesne programske opreme Slurm,
  • osnovno znanje operacijskega sistema Linux in lupine Bash
  • osnovno poznavanje programskega jezika Python

 

Ciljna publika: raziskovalci, inženirji, študenti, vsi ki potrebujejo več računskih virov pri svojem delu

Na izobraževanju pridobljena znanja:

  • Uporaba programskega jezika Python v superračunalniškem okolju
  • Uporaba knjižnic za vzporedno računanje v (RAPIDS, numba)
  • Uporaba knjižnice za porazdeljeno računanje DASK
  • Globoko učenje na superračunalniški gruči s knjižnico PyTorch

 

Organizator:

 

FRI logo

Predavatelji:

Ime:Davor Sluga
Opis:https://fri.uni-lj.si/sl/o-fakulteti/osebje/davor-sluga 
E-mail:davor.sluga@fri.uni-lj.si
Ime:Nejc Ilc
Opis:https://fri.uni-lj.si/en/about-faculty/employees/nejc-ilc
E-mail:nejc.ilc@fri.uni-lj.si

2nd Training School COST Action COSMIC WISPers (CA21106)

✇Indico Feed [Home]
10 June 2024 at 05:00

The European COST Action CA21106 "Cosmic WISPers in the Dark Universe: Theory, astrophysics and experiments" is an initiative funded by the COST Association, whose goal is to study axions and other very weakly interacting slim  particles (WISPs) emerging in several extensions of the Standard Model of Particle Physics. Searches for WISPs are strongly motivated by our attempts to understand the nature of the dark matter and puzzling astrophysical and particle physics observations.

The COST Action CA21106 2nd Training School will be held in Ljubljana (Slovenia). The goal is to start training a generation of young scientists in the interdisciplinary expertise on WISPs, ranging from theoretical and experimental  approaches to multi-messenger astroparticle physics and cosmology. For this purpose, the Training School will offer lectures from experts on different aspects of WISPs physics. The School is aimed at PhD students, as well as early-career postdocs.
 

[DBS seminar series] Bernat Corominas-Murtra, "Topological transitions and cell-cell adhesion properties in development"

✇Indico Feed [Home]
10 June 2024 at 10:00

The developmental process implies precise but significant changes in the geometry and structure of the embryonic tissues. Recent results show that apparently minor changes in adhesion at the local level of cell-cell contacts trigger deep, non-linear transitions in the topological organization of the whole embryonic tissue and, in consequence, change its material properties and facilitate geometrical deformations. Open questions remain, like the potential existence of causal feedbacks between cell differentiation processes and the potential heterogeneity of topological patterns existing within the tissue. The results presented in this talk are based on the fundamental assumption that predictive frameworks in theoretical biology must explore the  connection between different scales of the system. In this particular case, we use topology to establish a bridge between the system's scales defined by i) the cell level and ii) the tissue level. This enables us to predict global, non-trivial behaviours in tissues from the empirically feasible observables related to the mechanical properties of single cells.

Workshop: Python binding for C/C++

✇Indico Feed [Home]
10 June 2024 at 11:00

Description: Scientific and math software is written in C/C++ which often requires longer development time when combining a larger set of such libraries. To better utilize such libraries it is often convenient to interface them to a modern interpreted language such as Python.

We will learn how to bind C/C++ code in Python 3, starting by creating a Python object that will utilize one of the standard C/C++ containers, the vector. Continuing we will then take an existing C/C++ library and prepare it to be called from Python. Finally we will demonstrate a workflow where two different libraries are combined in a single Python 3 script.

Difficulty: Basic

Language: English

Prerequisite knowledge: Basic knowledge of Linux, the Terminal, and Python

Target audience: The workshop is intended for beginners and more experienced people interested in utilizing C/C++ code from Python

Skills to be gained:

  • Identifying useful libraries, and how to combine them in a workflow
  • Basics of libraries for Python
  • Binding C/C++ libraries in Python

 

Maximum number of participants: 15

Workflow: The training is live, in the afternoon. The workshop will combine lecture and practical parts, where your own laptop suffices (however, if need be we users can also make use of the NSC Cluster at the Jožef Stefan Institute (as an educational member of SLING)).

Location: Faculty of Mathematics and Physics (soba F2 - 1.nadstropje), Jadranska 19, 1000 Ljubljana

Organizer:

Univerza v Ljubljani v leto 2024 ...

Lecturer:

Name:Luka Leskovec
Description:Scientist and educationalist involved in theoretical physics and supercomputing
E-mail:luka.leskovec@fmf.uni-lj.si

[DBS seminar series] Tatyana Gavrilchenko, "Lessons from the tracheal terminal cell: what a unicellular network can teach us about distribution network design principles"

✇Indico Feed [Home]
13 June 2024 at 12:15
The insect respiratory system is a network of air-filled tubes permeating the animal body, supplying oxygen for metabolic activity and removing waste carbon dioxide. The majority of gas exchange occurs in the finest regions of the tracheal system, the terminal cells. These cells have a unique and highly specialized tree-like structure with long thin branches, reminiscent of neuronal arbors. While many aspects of the terminal cell are understood on a molecular level, including the mechanisms that guide branch extension and lumen formation, the macroscopic network features that allow for proper oxygen distribution remain mysterious. We use the Drosophila terminal cell as a model system for fundamental developmental problems of network structure and functions, utilizing an imaging data set that fully maps the structure of over one hundred individual cells.
 
First, we find that scaling relations succinctly encapsulate the dynamics of growing networks, and use the empirical scalings observed in the terminal cells to construct a minimal model of network growth that describes the system. Second, to understand the interplay between structure and function in the trees, we developed a model of oxygen distribution by a network embedded in a two-dimensional absorbing tissue, driven purely by diffusion along oxygen partial pressure gradients. Our method works well on complex geometries, including curved and branched networks that approximate the geometry of the terminal cells. These investigations offer insights into mammalian capillary networks, which have different structural features and delivery mechanisms from terminal cells but are guided by similar developmental principles. Understanding the salient features that govern the structure of these biological networks is essential to designing synthetic vasculature, a major step in the manufacture of artificial organs.

Ezequiel Alvarez: The Information frontier at the LHC

✇Indico Feed [Home]
17 June 2024 at 12:00

Once we have obtained the data, the question is how much information is encoded in it and how we can extract it. Any framework with this purpose utilizes information-besides-data and mathematical and statistical tools.  Bayesian statistics has the virtue that, by modeling the inner structure of the data, it allows us to inject prior information and its uncertainty as a catalyst to push out the information from the data.   The framework also possesses tools to assess the results unbiasedness and its consistency with the data.   We will show with examples at the LHC how to apply Bayesian tools in order to efficiently extract information about the physics inside.  We discuss possible applications in improving data-driven methods, and in pp > hh > bbbb, however the generality of the results apply to rethinking a variety of observables.  The method can exploit correlations at the event-by-event level and previously  untapped strengths such as for instance continuity and unimodality of some distributions.  We raise the question of whether there might be room for improvements in some of the current LHC analyses.

Course: MaX school: materials and molecular modelling with QUANTUM ESPRESSO

✇Indico Feed [Home]
19 June 2024 at 07:00

The course is organised by the MaX CoE, NCC Austria, NCC Czechia, NCC Hungary, NCC Poland, NCC Slovakia and NCC Slovenia

Description: The course encompasses a comprehensive curriculum designed to cover the primary features of the QUANTUM ESPRESSO code. The emphasis is on practical skill development. The course strikes a balance between theory and application, offering a hands-on learning experience. It caters to a beginner to intermediate level, aiming to equip participants with the fundamental knowledge and skills necessary for the effective utilization of QUANTUM ESPRESSO in their research and academic pursuits. 

Difficulty level: Beginner

Language: English

Prerequisite knowledge: Background in condensed matter physics or chemistry, basics of the LINUX shell

Skills to be gained:

The school is designed for participants with a background in condensed matter physics or chemistry interested in learning to use QUANTUM ESPRESSO. The school plans to cover the main features of the code and provide basic user skills such as compilation, simple scripting, choice of parallel options, and similar.

Location: Online (ZOOM) - Only registered participants will receive the Zoom link.

Lecturers:

Stefano de Gironcoli is a Professor of Computational Condensed Matter Physics at the International School for Advanced Studies (SISSA) in Trieste. Stefano obtained his Ph.D. at the Ecole Polytechnique Federale de Lausanne in 1992 with a thesis on “computational studies of substitutional alloys”, showcasing some of the early applications of Density Functional Perturbation Theory. He joined the SISSA faculty in 1994 as a Junior Research staff. A long-time developer and contributor to the Quantum ESPRESSO codes, he has been involved in many Summer Schools and tutorials about using the Quantum ESPRESSO suite. 

Pietro Delugas: is a Research Software Engineer at SISSA. Since 2015, he has been working on the development of QUANTUM ESPRESSO, collaborating with the MaX CoE and the QE Foundation. He works on the GPU porting, efficiency enhancement, and maintenance of the QE suite; he is also working on interoperability features of the suite, refactoring and updating the I/O, and developing postQE and qeschema Python packages. Pietro obtained his Ph.D. in physics at the University of Cagliari in 2006. He worked for Philips/NXP at the IMEC centre in Leuven, for the CNR at the IMM in Catania, and for the IOM in Cagliari before moving to SISSA in Trieste in 2015.

Oscar Baseggio: is a Research Software Engineer at SISSA. He has been working as a QE developer since 2018. In particular, he has focused on the maintenance and development of linear response codes and on the GPU porting. Oscar obtained his PhD in 2016 at the University of Trieste. He worked as an ADF developer for the SCM group in Amsterdam until 2017.

Anton Kokalj is a senior researcher at Jožef Stefan Institute (Ljubljana, Slovenia). He is active in computational chemistry and materials science and has expertise in modelling the physicochemical processes at surfaces relevant to corrosion and heterogeneous catalysis. He also explored other chemical systems, such as Li-battery materials, inorganic molecular crystals, organometallic complexes, and hydrogen-bonded systems. In the context of QUANTUM ESPRESSO, he has developed scripting and graphical interfaces, i.e., PWTK – PWscf ToolKit, PWgui – PWscf Graphical User Interface, and QE-modes.

Matic Poberžnik: is a postdoctoral researcher, working at the Jožef Stefan Institute, in Ljubljana, Slovenia. He is an expert in computational modelling with the Quantum ESPRESSO suite for electronic structure calculations and obtained his PhD in the computational studies of the oxidation of aluminium surfaces and their interactions with potential corrosion inhibitors. He is also a developer of the plugin ARTn, a new implementation of the Activation Relaxation Technique, which is an algorithm for identifying saddle points on potential energy surfaces. 

Laura Bellentani is an HPC Scientific Application Engineer working at the CINECA Supercomputing Center. She is committed to the GPU offload and optimization of materials-science codes in European projects, and she is highly motivated to work in the field of performance monitoring and profiling. Before joining the High-Level Support Team at CINECA, Laura obtained her PhD in Physics in 2019 at the University of Modena and Reggio Emilia and worked as a postdoc at CNR-NANO for IQubits.

❌
❌