Saturday, November 30, 2024

Notes of a performance on Water and Time that I gave sometime ago

 They told me that I should simply speak, without the aid of those tools and images I usually surround myself with. So, I’ll give it a try, speaking off the cuff and relying, forgive me, on some notes and a glass of water. Tracing back to the origin of water, we are drawn to the springs from which it flows. In reality, this bubbling forth is the result of an accumulation that pervades a portion of a mountain, a summary of underground stories and, in its most intense phases, even exposed to the sunlight.

This prehistory gives way to a constant and undeniable element of progress, from the source downwards. Now it is a stream, now a brook tumbling over rocks, now a river. With only minor diversions, there is but one direction—an accumulation. Time advances, entropy grows, aerial energy is both stored and dissipated. It’s not just water that flows; there’s sediment too, a piece of the mountain being carried toward the sea. Gutta cavat lapidem. Tectonic forces built it, and unyielding patience has worn it down.

Time accumulates, and there is a story that can only be remembered.

The above is the incipit of a performance-talk I gave a few years ago at MUSE. It is in Italian and you can find the notes here. 

Friday, November 8, 2024

How Generative AI Can Improve Hydrological Modeling and Support Scientists' Work

Generative artificial intelligence, with its ability to process and generate human-like text based on vast datasets, offers significant potential to enhance the interpretation of hydrological models, predict water cycle dynamics, and understand complex environmental interactions. By leveraging large language models (LLMs), researchers can optimize data processing, automate the extraction of relevant information from the extensive scientific literature, receive informed support in model programming, and generate descriptions of phenomena in natural language—extracting insights from vast datasets that go beyond the feasibility of manual analysis. Moreover, LLMs facilitate interdisciplinary collaboration by translating complex hydrological concepts into accessible language for diverse audiences. This seminar provides a brief introduction for selected high school students to generative model concepts, explores current applications, and examines the benefits and challenges of integrating LLMs in hydrology and related fields, highlighting their potential to advance scientific understanding and practical water resource management.

La presentazione in Italiano può essere trovata qui.