Idiap on LinkedIn Idiap youtube channel Idiap on Twitter Idiap on Facebook
Personal tools
You are here: Home Awards and Jobs News New opening for an internship position

New opening for an internship position

Executable Paper System for a Web-based Scientific Development Platform

The BEAT European Project built a web-based online platform ( https://www.beat-eu.org/platform/ ) for the development and benchmarking of scientific algorithms for applied pattern recognition and machine learning tasks.

Users define workflows and algorithms exclusively using a web-based system and schedule code to run on the BEAT computing farm with simple mouse clicks. The platform then presents results in the form of tables and figures that are programmed by the user. Results can be attached to scientific publications and certified by the platform. The BEAT platform is able to produce high-quality figures and graphics that could be included in scientific publications.

First objective

In this internship the master student will define, prototype and develop a web-based publication system that allows users of the BEAT platform to create high-quality scientific reports including tables and figures produced at the platform itself.

By clicking on those items, the user is re-directed to relevant information to reproduce claimed results.

This first objective is to provide a sort of executable scientific report system that could potentially replace currently existing publication vehicles such as pre-print servers, peer-reviewed conferences and journals, but which integrates access to scientific resources for the reproducibility of claimed results.

Second objective if time allows

The BEAT platform includes also the possibility for users to share code, toolchains and experimental results among pre-defined groups of users.

To transform the platform into a social scientific development system, it would be desirable to allow users to automatically discover new content (i.e. scientific resources) which are mature or provide satisfactory results in a variety of conditions. While some of this work can be done automatically, human annotation can be used to boost quality markings of algorithms and toolchains in a similar way which is done by social networks such as Facebook or Google+.

To motivate resource annotation by users, positive behavior shall be reinforced by user access to larger computing power.

The second objective of this internship is, if time allows, to gamify the BEAT platform so that high-quality content is brought forward and good work rewarded by giving priority to processing power on the computing farm.

To apply or to get more info about the position, click on the following link: Executable Paper System for a Web-based Scientific Development Platform

 

Dec 10, 2015
Document Actions