Decrypting the Language of Climate Change through Natural Language Processing machine learning Techniques

by Davide Faranda (CNRS Researcher in Climate Sciences)

Tuesday 30 May 2023, 15:00 → 16:00 Europe/Rome

Aula Magna (Dipartimento di Matematica)

In this lecture, I will be presenting exciting ways in which machine learning methods from Natural Language Processing (NLP) can be applied to the important issue of climate change. I will discuss the various techniques and algorithms used in NLP to process large amounts of data related to climate change. I will highlight the benefits of applying these methods, including the ability to extract important information from unstructured data, and to gain insights into the underlying patterns and trends in climate change data, with a focus on European weather extreme events such as heatwaves and extreme precipitation events. Furthermore I will discuss how mathematical models can be used to represent the linguistic structure of climate change text data, such as topic modeling and clustering. By combining these techniques with machine learning algorithms, we can discover hidden themes and topics, and even predict future trends in climate change. Overall, this lecture will demonstrate how machine learning and mathematics can be used in tandem to make sense of the vast amounts of data generated by climate change discourse, and how this knowledge can help us to better understand and address one of the most pressing challenges facing our planet today.