this is a collaborative effort and we are eager to get feedback from others (i.e. potential users and contributors): what else to integrate? how to improve text, design/IA, tutorials?
who is "we": a mixed team from CERN IT, Scientific Information Sevice and the LHC collaborations
Name: CERN Analysis Preservation
Description: A prototype service to preserve the insider knowledge about an analysis (data, software, docs), with easy open publishing options. Demo of latest prototype, concept and how it integrates with existing tools at CERN
Skills needed / sought: Some coaching on text similarity (I'm just using Jaccard similarity).
Additional info:
The Italian Senate is clogged by computer-generated amendments. This project aims to cluster similar amendments so that specific Senate procedures can be used to get rid of them in one sweep.
This is more of an open-politics project rather than open-science, but here it goes : )
Skills sought: experience with BVLC/caffe or will to reimplement 1409.7495 in a different framework
Additional Info:
Machine learning classifiers are often trained on simulated events and applied to real data. The authors of 1409.7495 deal with image recognition, trained with professional photos to be applied to low quality smartphone pictures. So … what they came up with should be applicable in physics.
As alternative to porting the physics data to caffe, one could as well port the method by 1409.7495 to theano/sklearn/tmva/whatever-works
Description: An open access and community driven collection of resources to help everyone get started on state of art Machine Learning Resources on open problems in Science !!
Skills Sought: Experience with Data Science and/or desire to teach and learn and/or knowledge about a bucket load of interesting open problems from your domain of expertise
Skills needed / sought: Experience with Deep learning techniques and performant numpy/scipy.
Name: luigi analysis workflow
Description: High-level extension layer for spotify's Luigi to enable advanced analysis workflows over distributes resources, so a WMS for end-user analysis =)
CERN Study Group Projects
List your public projects (with link to GitHub / Bitbucket / GitLab) below so we have a list for contributors. Format:
(N.B.: Do not use this issue for :+1: /
+1
and similar comments. Use it exclusively to list your open-science projects.)