MétaCan
Menu
Back to cohort
Record W6912680732 · doi:10.5281/zenodo.7461705

D4.1: Interim PRACE Training Report

2020· article· en· W6912680732 on OpenAlexaboutno aff

Bibliographic record

VenueZenodo (CERN European Organization for Nuclear Research) · 2020
Typearticle
Languageen
FieldComputer Science
TopicOnline Learning and Analytics
Canadian institutionsnot available
FundersHorizon 2020 Framework Programme
KeywordsDeliverableTraining (meteorology)InterimSummitWork (physics)Pandemic

Abstract

fetched live from OpenAlex

The Training Work Package (WP4) of the PRACE-6IP project is responsible for the design and execution of a comprehensive range of HPC training activities. This includes an extensive, annual programme of short courses that are delivered by a network of now 14 PRACE Training Centres (PTCs), a series of Seasonal Schools, on-demand events, Massive Open Online Courses (MOOCs) which are underpinned by the PRACE Training Portal. This deliverable is the interim, mid-term report of all PRACE-6IP training activities up to Month 18 of the project. The COVID-19 pandemic has had an obvious effect on training activities, in particular face-to-face courses and schools that are no longer feasible due to travel and social distancing restrictions. While this had caused disruptions and the postponement and cancellations of some 20+ events (PTC courses, a Seasonal School, the 2019 International HPC Summer School, the EuroHPC Summit Week 2020 hands-on workshops), there has been a quick pivot to providing courses online. Since the onset of the pandemic in March 2019, WP4 had still managed to deliver 29 PTC courses and the PRACE Autumn School 2020 online, not the mention converting the PRACE Summer of HPC into a virtual mentoring programme and continuing the Massive Open Online Courses (MOOCs). The network of PTCs has expanded by four new members (from Austria, Belgium, Slovenia and Sweden) in late 2019 for a total of 14 partners who are involved in delivering an annual joint programme of courses. From May 2019 to September 2020, the PTCs have collectively delivered 119 courses, representing 326 days of training with 3,549 participants. This includes the 2019-2020 PTC programme in which almost a quarter of its courses had to be delivered online. As trainers continue to learn and improve the online learning experience, indications are that online courses have proven to be attractive to participants from all over Europe including non-PTC hosting countries. These lessons will be brought into the 2020-2021 programme with a target of 138 courses, representing 338 training days. WP4 had targeted the delivery of eight Seasonal Schools during the course of the PRACE-6IP project. The first PRACE Autumn School in 2019 was held in Slovenia organised in a fast-tracked manner. While the schedule of Seasonal Schools, including location and organising partners, were finalised by early 2020, the COVID-19 pandemic caused the 2020 PRACE Winter School in Austria to be aborted after its first day. The subsequent 2020 Autumn School in Slovenia was successfully delivered in a hybrid face-to-face and online format. Therefore, two Seasonal Schools have been fully delivered with a total of 128 participants. A modified schedule is being finalised for delivery of remaining Seasonal Schools in 2021. Apart from Seasonal Schools, two on-demand events have been held in 2019 that involved collaborations with the EUDAT and the BioExcel CoE projects. Meanwhile, WP4 has co-organised another successful International HPC Summer School 2019, in collaboration with partners from Canada, Japan and the U.S., which was held in Kobe, Japan in July 2019, with 80 participants including 30 from Europe. The 2020 International HPC Summer School was planned to take place in Toronto, July 2020 but has been cancelled due to the COVID-19 pandemic; the event is being rescheduled to 2021 in the same venue. Plans are ongoing to organise an advanced PRACE extreme-scale workshop, taught by a pool of instructors from PRACE partners, in 2021. The PRACE Summer of HPC programme, where European students are mentored by HPC experts across participating PRACE partner sites, hosted 25 projects for 25 visiting students at HPC centres in 2019. While travel restrictions in summer 2020 had looked to curtail this programme in 2020, WP4 managed to transform it into a remote mentoring programme where 50 students partnered up to partake in 25 projects across HPC partners. With the addition of another new MOOC developed by WP4, PRACE now has a repertoire of six MOOCs that are being run on the FutureLearn platform on a periodic basis. During the first half of the project, 5 MOOCs were delivered across 8 instances (i.e. some repeat runs of the same MOOC), which had attracted a total of 7,475 active learners (16,568 enrolments). Development of new MOOCs are planned for the second half of the project. Training collaboration has always been an important aspect of PRACE. WP4 has engaged with a variety of external projects and research infrastructures to identify potential synergistic activities. Some of the engagements so far include the planning of joint training event with RedCLARA (South America), potential joint activities with CERN / GÉANT / SKA and working with FocusCoE and CASTIEL to ensure alignment in activities and providing a platform for exchange of ideas and best practices. New approaches to develop basic training materials are being investigated with the intention to contribute towards the collaborative efforts of HPC Carpentry. Finally, the PRACE Training Portal and Events Portal have been updated and improved with new functionalities to improve the user experience. The Training Portal underwent a redesign to be consistent with the main PRACE web site, with the adoption of a new events registry system with additional metadata that makes it easier for users to identify relevant courses.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.011
metaresearch head score (Gemma)0.016
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesInsufficient payload (model declined to judge)
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Not applicable · Consensus signal: Not applicable
GenreCandidate signal: Other · Consensus signal: Other
Teacher disagreement score0.318
Threshold uncertainty score0.973

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0110.016
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0000.001
Bibliometrics0.0030.002
Science and technology studies0.0020.001
Scholarly communication0.0050.003
Open science0.0030.005
Research integrity0.0020.003
Insufficient payload (model declined to judge)0.3180.240

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.092
GPT teacher head0.292
Teacher spread0.200 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

Study designNot applicable
Domainnot available
GenreOther

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2020
Admission routes1
Has abstractyes

Explore more

Same venueZenodo (CERN European Organization for Nuclear Research)Same topicOnline Learning and AnalyticsFrench-language works237,207