The Orion Star-Formation Complex as a training set for machine learning techniques

As the field of star-formation follows astronomy into the era of big data, we are now faced with the challenges behind truly reliable and unbiased applications of artificial intelligence and machine learning techniques. One of these challenges is quantifying the uncertainties behind machine learning applications and evaluating how the uncertainties in our prior knowledge of the field of starformation impact the performance of such algorithms. The Orion star-formation complex is arguably the most well-studied star-forming region in the sky. With more than 8,000 refereed publications indexed by NASA/ADS and about one-tenth of those feeding data products to the CDS database, the region offers a wealth of possible attributes to be explored by machine learning techniques. In this study, we examine the region under the framework of the NEMESIS (New Evolutionary Model for Early stages of Stars with Intelligent Systems) project - which aims to reshape our understanding of star-formation by employing artificial intelligence methods to explore astronomical big-data. Towards benchmarking the Orion complex as a training set for machine learning techniques, we present a synoptic catalogue of Young Stellar Objects in the region. This catalogue combines data from large surveys (e.g., AllWISE, AKARI, Spitzer, Herschel, Gaia EDR3, etc.) with literature data 109 products, covering the whole electromagnetic spectrum from X-rays to radio and has been curated with particular attention to both time-domain and variability aspects.