Finding patterns in planets

Elizabeth Tasker
宇宙航空研究開発機構 宇宙科学研究所

In the last twenty years, we have discovered around 3,500 exoplanets. As new missions threaten to explode that number to even greater proportions, the challenge becomes how we analyse this data. Neural networks offer a potential way to find patterns within the exoplanet properties that allow planets to be categorised more successfully that can be done by eye. Such a technique has the advantage of handling large datasets and also the potential to analysis the data without using preconceived ideas about planet populations based on existing formation theories. This talk will discuss how successful such an approach has been.

Place: 2F Conf. room / 研究管理棟２階会議場（1236号室）