Скачать с ютуб Lasse Espeholt - Weather Forecasting using Deep Learning - A paradigm shift (MetNet-3) в хорошем качестве

Lasse Espeholt - Weather Forecasting using Deep Learning - A paradigm shift (MetNet-3)

Talk recorded on June 6th 2023 Lasse Espeholt from Google DeepMind presented his work on MetNet-3. Enhancing the accuracy and reliability of weather forecasting is of paramount importance due to its significant implications on society, the economy, and the environment. Accurate weather predictions are crucial for several sectors including agriculture, energy, aviation, and disaster management, among others. Improved forecasts enable proactive planning and risk mitigation, thus potentially saving lives and reducing economic losses. With advancements in technology, particularly in the areas of machine learning and artificial intelligence, we have the opportunity to revolutionise traditional weather forecasting methodologies. In this talk, I will present our past work on neural network-based weather forecasting that outperforms traditional approaches, and what these neural network-based approaches enable.