Focus on: 
Back to Conference View
Login

Inverted CERN School of Computing 2017

→ Europe/Zurich
31/3-004 - IT Amphitheatre (CERN)

31/3-004 - IT Amphitheatre

CERN

105
Show room on map
Description

The Inverted CERN School of Computing consists of lectures presented over a few days by former CERN School of Computing students, providing advanced education in specialist topics. The Inverted School provides a platform to share their knowledge by turning the students into teachers.

Now in its 10th year, we welcome you to the iCSC 2017!

Anyone can attend - it's absolutely free*.  We also webcast the event so that you can tune in from anywhere in the world.

If you register, we will provide you the iCSC 2017 booklet (while stocks last!) which contains all the slide materials and speaker information, as a hard-copy reference and souvenir. We'll also make sure enough tea/coffee is available if we know you're coming.

More information on the Inverted School events can be found at http://csc.web.cern.ch/inverted-school

* We offer free lectures, a free booklet, a wonderful recorded webcast and free tea/coffee for you to enjoy during the breaks where you can discuss with the speakers and other students. You are responsible for your own travel, visa, food and accommodation if you wish to attend as a member of the audience.

iCSC2017_booklet.pdf
Webcast
There is a live webcast for this event
Watch
Contact
  • Monday, 6 March
    • 14:00 → 14:15
      A word from the IT Department Head 15m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
      Speaker: Frederic Hemmer (CERN)
      Recording
    • 14:15 → 14:30
      Introduction to the inverted CSC 15m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
      Speaker: Sebastian Lopienski (CERN)
      2017.03.06 iCSC-2017 Welcome.pdf
      2017.03.06 iCSC-2017 Welcome.pptx
      Recording
    • 14:30 → 15:30
      Let your machine do the learning - Lecture 1 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      The field of Artificial Intelligence, whose formal definitions go as back as the 40's, have recently gained a renowned interest in the community as more and more problems become amenable to be tackled by it. Even better, we are now available to try complex techniques in a very accessible manner, lowering the entrance admission to this cool club to just a couple of hours.

      In this series of lectures, we are going to explore non-conventional techniques to solve long standing problems, coming from AI, from a pragmatic and up to date perspective. By the end of these lectures you should be able to get your hands dirty with exciting real examples. You should be able to identify what kind of problems are you dealing with, what tools does AI have in store for you and how to apply them in a straightforward way, with room for depth. Just enjoy longer coffee breaks as the machine works it out for you.

      Speaker: Daniel Hugo Campora Perez (CERN & Universidad de Sevilla)
      let_your_machine_do_the_learning_1.pdf
      Recording
    • 15:30 → 16:00
      Coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 16:00 → 17:00
      Algorithms for Anomaly Detection - Lecture 1 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      The concept of statistical anomalies, or outliers, has fascinated experimentalists since the earliest attempts to interpret data. We want to know why some data points don’t seem to belong with the others: perhaps we want to eliminate spurious or unrepresentative data from our model. Or, the anomalies themselves may be what we are interested in: an outlier could represent the symptom of a disease, an attack on a computer network, a scientific discovery, or even an unfaithful partner.

      We start with some general considerations, such as the relationship between clustering and anomaly detection, the choice between supervised and unsupervised methods, and the difference between global and local anomalies. Then we will survey the most representative anomaly detection algorithms, highlighting what kind of data each approach is best suited to, and discussing their limitations. We will finish with a discussion of the difficulties of anomaly detection in high-dimensional data and some new directions for anomaly detection research.

      Speaker: Michael Davis (CERN)
      AlgorithmsForAnomalyDetection_Part1.pdf
      Recording
  • Tuesday, 7 March
    • 08:30 → 09:00
      Welcome coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 09:00 → 10:00
      Applying natural evolution for solving computational problems - Lecture 1 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      Darwin’s natural evolution theory has inspired computer scientists for solving computational problems. In a similar way to how humans and animals have evolved along millions of years, computational problems can be solved by evolving a population of solutions through generations until a good solution is found.

      In the first lecture, the fundaments of evolutionary computing (EC) will be described, covering the different phases that the evolutionary process implies. ECJ, a framework for researching in such field, will be also explained.

      In the second lecture, genetic programming (GP) will be covered. GP is a sub-field of EC where solutions are actual computational programs represented by trees. Bloat control and distributed evaluation will be introduced.

      Speaker: Daniel Lanza Garcia (CERN, Switzerland)
      Lecture 1.pdf
      Lecture 1.pptx
      Recording
    • 10:00 → 11:00
      Distributed consensus and fault tolerance - Lecture 1 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      In a world where clusters with thousands of nodes are becoming commonplace, we are often faced with the task of having them coordinate and share state. As the number of machines goes up, so does the probability that something goes wrong: a node could temporarily lose connectivity, crash because of some race condition, or have its hard drive fail.

      What are the challenges when designing fault-tolerant distributed systems, where a cluster is able to survive the loss of individual nodes? In this lecture, we will discuss some basics on this topic (consistency models, CAP theorem, failure modes, byzantine faults), detail the raft consensus algorithm, and showcase an interesting example of a highly resilient distributed system, bitcoin.

      Speaker: Georgios Bitzes (CERN)
      Distributed consensus and fault tolerance - part 1.pdf
      Distributed consensus and fault tolerance - part 1.pptx
      Recording
    • 11:00 → 11:30
      Coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 11:30 → 12:30
      Algorithms for Anomaly Detection - Lecture 2 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      The concept of statistical anomalies, or outliers, has fascinated experimentalists since the earliest attempts to interpret data. We want to know why some data points don’t seem to belong with the others: perhaps we want to eliminate spurious or unrepresentative data from our model. Or, the anomalies themselves may be what we are interested in: an outlier could represent the symptom of a disease, an attack on a computer network, a scientific discovery, or even an unfaithful partner.

      We start with some general considerations, such as the relationship between clustering and anomaly detection, the choice between supervised and unsupervised methods, and the difference between global and local anomalies. Then we will survey the most representative anomaly detection algorithms, highlighting what kind of data each approach is best suited to, and discussing their limitations. We will finish with a discussion of the difficulties of anomaly detection in high-dimensional data and some new directions for anomaly detection research.

      Speaker: Michael Davis (CERN)
      AlgorithmsForAnomalyDetection_Part2.pdf
      Recording
    • 12:30 → 14:00
      Lunch 1h 30m
    • 14:00 → 15:00
      Virtual Machine Images Management in Cloud Environments 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      Nowadays, the demand for scalability in distributed systems has led a design philosophy in which virtual resources need to be configured in a flexible way to provide services to a large number of users. The configuration and management of such an architecture is challenging (e.g.: 100,000 compute cores on the private cloud together with thousands of cores on external cloud resources). There is the need to process CPU intensive work whilst ensuring that the resources are shared fairly between different users of the system, and guarantee that all nodes are up to date with new images containing the latest software configurations. Different types of automated systems can be used to facilitate the orchestration. CERN’s current system, composed of different technologies such as OpenStack, Packer, Puppet, Rundeck and Docker will be introduced and explained, together with the process used to create new Virtual Machines images at CERN.

      Speaker: Lorena Lobato Pardavila (CERN & Universidad de Oviedo)
      iCSC_VMClouds.pdf
      iCSC_VMClouds.pptx
      Recording
    • 15:00 → 15:30
      Coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 15:30 → 16:30
      Let your machine do the learning - Lecture 2 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      The field of Artificial Intelligence, whose formal definitions go as back as the 40's, have recently gained a renowned interest in the community as more and more problems become amenable to be tackled by it. Even better, we are now available to try complex techniques in a very accessible manner, lowering the entrance admission to this cool club to just a couple of hours.

      In this series of lectures, we are going to explore non-conventional techniques to solve long standing problems, coming from AI, from a pragmatic and up to date perspective. By the end of these lectures you should be able to get your hands dirty with exciting real examples. You should be able to identify what kind of problems are you dealing with, what tools does AI have in store for you and how to apply them in a straightforward way, with room for depth. Just enjoy longer coffee breaks as the machine works it out for you.

      Speaker: Daniel Hugo Campora Perez (CERN & Universidad de Sevilla)
      let_your_machine_do_the_learning_2.pdf
      Recording
  • Wednesday, 8 March
    • 08:30 → 09:00
      Welcome coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 09:00 → 10:00
      Creating Effective Data Visualizations - Lecture 1 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      In this course I aim to give an overview of data visualisation as a field, including many of the important theoretical groundings in data visualization.

      We will explore the different ways of representing visual information, and the strengths/weaknesses of those approaches.

      Using real-world case studies, I will demonstrate techniques and best practices for visualizing complex multi-dimensional data common to high energy physics and other fields.

      Speaker: Eamonn Maguire (Pictet Asset Management)
      CERN-Vis-Course-Pt1.pdf
      Recording
    • 10:00 → 11:00
      Distributed consensus and fault tolerance - Lecture 2 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      In a world where clusters with thousands of nodes are becoming commonplace, we are often faced with the task of having them coordinate and share state. As the number of machines goes up, so does the probability that something goes wrong: a node could temporarily lose connectivity, crash because of some race condition, or have its hard drive fail.

      What are the challenges when designing fault-tolerant distributed systems, where a cluster is able to survive the loss of individual nodes? In this lecture, we will discuss some basics on this topic (consistency models, CAP theorem, failure modes, byzantine faults), detail the raft consensus algorithm, and showcase an interesting example of a highly resilient distributed system, bitcoin.

      Speaker: Georgios Bitzes (CERN)
      Distributed consensus and fault tolerance - part 2.pdf
      Distributed consensus and fault tolerance - part 2.pptx
      errata.txt
      Recording
    • 11:00 → 11:30
      Coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 11:30 → 12:30
      Applying natural evolution for solving computational problems - Lecture 2 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      Darwin’s natural evolution theory has inspired computer scientists for solving computational problems. In a similar way to how humans and animals have evolved along millions of years, computational problems can be solved by evolving a population of solutions through generations until a good solution is found.

      In the first lecture, the fundaments of evolutionary computing (EC) will be described, covering the different phases that the evolutionary process implies. ECJ, a framework for researching in such field, will be also explained.

      In the second lecture, genetic programming (GP) will be covered. GP is a sub-field of EC where solutions are actual computational programs represented by trees. Bloat control and distributed evaluation will be introduced.

      Speaker: Daniel Lanza Garcia (CERN, Switzerland)
      Lecture 2.pdf
      Lecture 2.pptx
      Recording
    • 12:30 → 14:00
      Lunch 1h 30m
    • 14:00 → 15:00
      Creating Effective Data Visualizations - Lecture 2 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map

      In this course I aim to give an overview of data visualisation as a field, including many of the important theoretical groundings in data visualization.

      We will explore the different ways of representing visual information, and the strengths/weaknesses of those approaches.

      Using real-world case studies, I will demonstrate techniques and best practices for visualizing complex multi-dimensional data common to high energy physics and other fields.

      Speaker: Eamonn Maguire (Pictet Asset Management)
      CERN-Vis-Course-Pt2.pdf
      Recording
    • 15:00 → 15:30
      Coffee 30m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
    • 15:30 → 16:30
      Let your machine do the learning - hands-on session 1h 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
      Speaker: Daniel Hugo Campora Perez (CERN & Universidad de Sevilla)
      1. Log in to cernbox
      2. Import lab exercises on Swan
    • 16:30 → 16:45
      Closing remarks 15m 31/3-004 - IT Amphitheatre

      31/3-004 - IT Amphitheatre

      CERN

      105
      Show room on map
      Speaker: Sebastian Lopienski (CERN)
      2017.03.08 iCSC-2017 Closing.pdf
      2017.03.08 iCSC-2017 Closing.pptx
      Recording
Powered by Indico v3.3.2-pre