Noisy and/or Massive Datasets

As the complexity of the system is increasing concomitantly with the sophistication of the experimental investigation procedure, the development of appropriate and robust tools to interpret the experimental dataset is a crucial need. Characterization of complex systems often implies the acquisition of massive and/or noisy dataset. This frequently leads to a major trouble when trying to extract relevant information since data analysis tools available do not take into account the full complexity of the studied systems. As data analysis is an essential link between modeling and experimental measurement, one needs to optimize the extraction of the significant information from the dataset without losing and/or biasing it.

What we need:

• Extract signal from the noise.
• Analysis of massive dataset.
• How to deal with biased database ?
• Robustness of the method (limit)
• How to deal with black boxes ? (trying to become independent of the constraint imposed by the instruments suppliers).
• Validation of the model

→ need to develop custom-made and adapted tools to deal with all of the above problems.