First CERN openlab summer student lightning talk session
On Tuesday 13 August and Thursday 15 August, the 2019 CERN openlab summer students will present their work at a dedicated public Lighting Talk sessions.
In a 5-minute presentation, each student will introduce the audience to their project, explain the technical challenges they have faced and describe the results of what they have been working on for the past few weeks.
It will be a great opportunity for the students to showcase the progress they have made so far and for the audience to be informed about various information-technology projects, the solutions that the students have come up with and the potential future challenges they have identified.
-
-
1
Welcome and introduction
-
2
Real Time Server Monitoring and Fixed Point Inference in FPGA
The project dealt with real time monitoring of the server hosting the FPGA for Convolution Neural Network inference for the Proto-DUNE Project. I also investigated methods to deal with the numerical overflow of the weights and activation while working with fixed point arithmetic in FPGA.
Speaker: Debdeep Paul -
3
EOS Winston: Expert Systems for Automated Diagnosis and Remediation
The IT-ST group at CERN runs and evaluates innovative cloud storage technologies for their application to big data problems in high-energy physics research. One of the entities it focuses on is EOS, the CERN multi-Petabyte disk-based storage service built from commodity hardware, heavily used as well by LHC and non-LHC experiments. The massive scale at which EOS runs leads to room for multiple issues and anomalies to creep in. These need to be dealt with in real-time to ensure smooth operations.
The project aims to improve the current troubleshooting and diagnosis of the different components that compose the EOS infrastructure with the development of an Expert System that collects diagnostic information such as metrics, signals, and alerts, from each of the namespaces and assists engineers in reducing the time to debug these issues on the system, thus automating some of the troubleshooting tasks.
Speaker: Ishank Arora -
4
Improvements on BioDynaMo build systemSpeaker: Giovanni De Toni
-
5
Scaling-out CERN’s IDS System with Network Automation
CERN is one of the most heterogeneous network in the world and in order to keep its traffic safe we’ve to inspect it in real time.
We have already an Intrusion Detection System that inspect CERN traffic firewall, but we're looking to make it more powerfull and more reliable.
In my work I've focused on IDS upgrade in order to support multiple hardware vendors to improve its scalability and availability.
Speaker: Andrea Lacava -
6
Deep Graph Neural Networks for Fast HGCAL SimulationSpeaker: Raghav Kansal
-
7
EOS Integration into OpenStack Manila
OpenStack is a popular open source cloud-computing software platform used widely at CERN. EOS is a disk-based, low-latency storage service powering user, project, and experiment data on services such as CERNBox.
This project strives to improve user experience by integrating EOS into OpenStack Manila, a shared storage system. This way, users are able to request and access project space via a single interface.
Speaker: Ms Elisabeth Ann Petit-Bois (Kennesaw State University) -
8
Web Services - Discourse Forum AutomationSpeaker: Rajula Vineet Reddy
- 9
-
15:06
Break
-
10
Continuous integration for containerized scientific workflowsSpeaker: Leticia Farias Wanderley
- 11
-
12
Analysis and Modeling of Storage Access Patterns and Caching StrategiesSpeaker: Shreya Krishnan
-
13
Deep I/O performance analysis of CVMFS using modern Linux toolsSpeaker: Shahnur Isgandarli
-
14
DL Models for Particle Identification and Energy Regression in CMS HGCAL L1T
Using Micron's FPGA-based inference engines and FWDNXT firmware + software for compiling models and running the inference
Speaker: Anwesha Bhattacharya -
15
Fast Inference on FPGAs for HEP trigger systems
In order to benefit of modern machine learning in the early stages of the data acquisition of a typical HEP experiment, one has to be able to execute ML model inference within the
latency of the L1 trigger system. At the LHC, this time is of O(10) μs. The aim of this project is to deploy a set of LHC-related neural networks to the Intel FPGAs.Speaker: Hamza Javed -
16
Benchmarking tools for NextGen Archiver for WinCC OASpeaker: Jayaditya Gupta
-
17
The usage of deep learning methods for raw data preprocessing at protoDUNE experimentSpeaker: Maksim Artemev
-
18
Automation Tools for Invenio
Tool that generates report on the status of inveniosoftware repositories and suggests suitable actions.
Speaker: Foteini Panagiotidou -
16:29
Deliberations of the Jury
Results will be made available during the Lightning talk drink on Aug 15th.
-
1