Nicholas Clinton

We'll demystify the rapidly evolving landscape of AI in this survey tailored for the Earth Engine Partner. We'll survey Google’s AI tools for geospatial and Earth AI—including Earth Engine & AlphaEarth Foundations (AEF), Geospatial Reasoning, Gemini, Spark, and Co-Scientist.

Many would say that Land Classification is Earth Engine's raison d'etre. Having built a classifier, new users of Earth Engine are often amazed at their ability to pan and zoom across the entire earth's surface and comprehensively see the results of their model.

Numerous publicly available Earth Observation maps provide data on forests, land use, and tree cover. However, these maps often differ because they use various definitions and classification systems.

The EU Deforestation Regulation aims to address the environmental and social issues associated with deforestation, specifically related to seven agricultural commodities. In this session, we’ll describe these commodities, showcase the state of the art in their detection and monitoring, and discuss the challenges and gaps that exist today.

Learn how to develop and train basic neural networks in TensorFlow, for use with Earth Engine. This session will cover vector inputs, suitable for tabular or per-pixel image models.

Building on NN1, this session will demonstrate how to use Apache Beam to build large training datasets, training a convolutional model of land cover on Vertex AI, deploying the model and visualizing predictions in Earth Engine.

In this breakout session, we will introduce attendees to the multimedia effort Cloud-Based Remote Sensing with Google Earth Engine: Fundamentals and Applications. This is the product of more than a year of volunteer effort from more than a hundred individuals, working in concert to provide coordinated open-access learning resources for learning how to use Earth Engine for the public good.

Get hands-on with ML in Earth Engine! This session is an end-to-end walkthrough of generating training and validation data in Earth Engine, exporting to the TFRecord format, building a simple DNN model in TensorFlow, training and evaluation with the Earth Engine data, classification of an image as a TFRecord file, and visualization in Earth Engine.

In this session, you'll learn all the ways you can reduce images, tables, image collections, lists, dictionaries, and other stuff. Also, learn everything else you ever wanted to know about reducers (but were afraid to ask).