Refining SED Data Configurations

Case study:

Optimizing Planet Detection Experiments

Delivered with:
Find out how

The ask

A European space agency research group was looking for new ways to improve their operational performance and grow new opportunities for investigation. In their research processes, the client uses Spectral Energy Distribution (SED) to analyze proto-solar systems, but their existing systems were taking too long to produce the results they required. 


What We Did

Using SED data from past experiments and a deep learning approach, we matched each SED from the dataset with their experimental configurations. The A.I. solution learned to relate the optimal experiment configuration with the desired SED. This relationship accelerated the process and enabled the client to execute a higher volume of valuable experiments in much less time. 


The Process

A huge number of experiments were executed in order to create a dataset that related to the experimental parameter configurations with the SED data.

Due to the nature of the data, the chosen A.I. solution was a deep learning approach. More specifically, the deep learning architecture was a feed-forward neural network trained with the previous dataset to learn how to match the configuration parameters with the SED data.

Using this trained model, we can now find out the best configuration of parameters needed in order to start an experiment.



Our solution was able to reduce the computing time of SED from an average of 52 hours to only 5 hours. This meant that our client could test up to ten times the number of models and get valuable results within the same amount of time. This greatly helped to increase productivity and accelerated the research progress significantly. 

The solution we created used a feed-forward neural network to optimize the execution of processes. Other scenarios where it could be applied include experiments or simulations in many industries where time is critical to obtain more outcomes in a shorter amount of time.


For a free assessment, answers to your questions, or to discover how software and AI can be integrated into your business, don’t hesitate to get in touch.