Post-Doctoral Research Visit F/M Reproducible Science and Software Variability

Contract type :
Fixed-term contract

Level of qualifications required :
PhD or equivalent

Fonction :
Post-Doctoral Research Visit

About the research centre or Inria department

The Inria Rennes – Bretagne Atlantique Centre is one of Inria’s eight centres and has more than thirty research teams. The Inria Center is a major and recognized player in the field of digital sciences. It is at the heart of a rich R&D and innovation ecosystem: highly innovative PMEs, large industrial groups, competitiveness clusters, research and higher education players, laboratories of excellence, technological research institute, etc.

Context

The work will be realized in the context of the RESIST_EA associate team, which gathers the DiverSE Inria

research team together with the VIAS and COMPLEX Departments of Simula Research Laboratory (Norway). The
DiverSE team is located in Rennes, Brittany, France at the Inria research center (Centre Inria de l’Université de
Rennes). DiverSE’s research is in the area of software engineering and variability, while VIAS is dedicated to the
validation of autonomous systems. The candidate will be employed by Inria and will work under the co-
supervision of Prof. Mathieu Acher (INSA Rennes, DiverSE team), Assoc. Prof. Paul Temple (Univ Rennes, DiverSE
team), Research Prof. Arnaud Gotlieb (Simula Research Laboratory, Norway) and Research Scientist Helge
Spieker (Simula Research Laboratory).

Every year Inria International Relations Department has a few postdoctoral positions in order to support Inria
international collaborations. Hence, the post-doc is part of DRI campaign.
This year, postdoctoral positions within the frame of Inria London, Inria Brasil and Inria Chile programs and to
strengthen partnerships with Simula (Norway), University of Waterloo (Canada) and KAIST and ETRI (South
Korea) are eligible. 

The postdoc contract will have a duration of 24 months. The default start date is November 1st, 2023 and
not later than January, 1st 2024. The postdoctoral fellow will be recruited by one of the Inria Centers in France
but it is recommended that the time is shared between France and the partner’s country (please note that the
postdoctoral fellow has to start his/her contract being in France and that the visits have to respect Inria rules for
missions).

Assignment

Software resilience is about studying the varying conditions under which software systems can resist to failures.
Thus, in order to ensure resilience, it is crucial to demonstrate that results produced by systems can be safely
and accurately reproduced. Usual experimental sciences like biology, medicine, chemistry and physics have built
strong procedures for such purposes, but still, as they are more and more guided by data processing and analyses,
there is a pressing need to revise such procedures to ensure better software reproducibility.
For instance, studies about climate change require the design of mathematical models, the mining and analysis
of data, execution of large simulations which involve computational tasks with specific software configurations
and, unfortunately, different configurations can lead to different and inconsistent results., It is not an
overstatement to say that computational science depends on software and its engineering.
The quest for reproducibility takes different forms and requires to make all data and code available in such a way
that the computations can be executed again with identical results. For achieving reproducible science, there areseveral challenges ahead involving multi-disciplinary collaboration and socio-technical innovation within
software development.

The software reproducibility problem can be captured by software variability modelling: many factors (operating
system, third-party libraries, versions, workloads, compile-time options and flags, etc) themselves subject to
variability can alter the results, up to the point it can dramatically change the conclusions of some scientific
studies. The goal of this post-doc is to make reproducible science more resilient through the management of software
variability.

There are two grand directions to consider. First, software variability can be accidental: a change in some
software parameters or libraries versions may incidentally change the conclusions. From this perspective,
methods and techniques should be developed to control that the result can still be computed and is still
functionally valid and coherent with regard to the original result.
Second, software variability can also be an opportunity to generate and explore new hypotheses of a scientific
study or rely on different methods or process different input data. From this perspective, methods and
techniques should be developed to automatically synthesize some diversity that is both coherent wrt software
and expert knowledge while covering a wide range of possibilities. This diversity should either exhibit new results
that contradict original results or make more robust the original studies.

We plan to consider at least the following case studies, spanning different fields and research expertise of DiverSE
and Simula: floating point computations, automated driving, cardiac modelling, and neuroimaging (eg https://www.narps.info/ ).

Main activities

Main activities (5 maximum) : The goal of this post-doc is to make reproducible science more resilient through the management of software variability. We plan to consider at least the following case studies, spanning different fields and research expertise of DiverSE and Simula: floating point computations, automated driving, cardiac modelling, and neuroimaging (eg https://www.narps.info/).

Planned visits at Simula (Norway): The candidate will be invited to visit Simula Research Laboratory in Norway not only to
interact with the research scientists working in the VIAS department, but also to get involved with other
application fields of software reproducibility such as scientific computing in cardiac modelling and intelligent
transport systems. In additional to remote, weekly meetings with both Inria and Simula partners, at least two
visits per year will be planned ahead in Norway.

Skills

Candidates must have completed their PhD before the starting date and should have a solid background in
one or multiple of the following areas:

• Automated software engineering (automated testing, modelling, etc.)
• Machine learning (statistical machine learning, large language models)
• Computational science (e.g., reproducible science)

The candidate should further have:

• An excellent publication track record in international conferences or journals
• Strong interpersonal skills and the ability to work and communicate well in an internationally
  collaborative environment

Possible and temporary derogations can be granted to exceptional candidates not having yet defended their PhD
but expecting to do so before the end of the 2023 year.

Benefits package

• Subsidized meals
• Partial reimbursement of public transport costs
• Possibility of teleworking (90 days per year) and flexible organization of working hours
• Partial payment of insurance costs

Remuneration

monthly gross salary amounting to 2746 euros