heading 1

heading 2

heading 3

heading 4

heading 5
heading 6

quote

Aicacia
Microlab Leader: Sherin Thomas ☉
Simon Benigeri ❍ Vladislav Gerasimov ❍ Simone Blanco Malerba ❍ Guru Prakash ❍ Jason Prasad ❍ Rahul Varma

Vision: Support global reforestation goals with a tool to gather and synthesize available reforestation knowledge.

Problem Statement: One way to support global reforestation goals is to ensure practitioners have ready access to useful information on proven best practices and techniques, as well as knowledge and resources specific to the geographies and landscapes in which they are working. There is a wealth of high-quality knowledge on reforestation, but due to the volume and technical nature of resources, sifting through available material to find relevant and actionable information for a particular project can be a barrier to progress. Our aim is to support practitioners by leveraging automation, data, and AI-driven tools to address two key aims:

1. Catalog, organize, and present repositories of knowledge in a user-friendly way
2. Provide search functionality and a Q&A platform to facilitate finding the information relevant to a particular project

A central part of our work involves training an AI model to address these challenges. While there are off the shelf tools available, these fall short because they are not trained on domain-specific information needed to understand the nuances of ecological regeneration. We will develop scalable approaches for generating training datasets from high-quality sources of data curated by those with deep knowledge in the reforestation field, with the intention of making these domain-specific models and datasets freely available to the community.

Program Phases: The Aicacia MicroLab will be conducted in three individual phases which will all come together to address the problem described above:

Phase 1: MVP - Build a search tool for reforestation material

Work involved:
1. Scrape data from open source journals and other publicly accessible resources
2. Process and store data in an online vector databaseUsing an off-the-shelf embedding model, build search and retrieval on top of (2) using the retrieval augmented generation (RAG) method.
3. Build a rudimentary search interface using tools such as Retool or Streamlit
4. Build a simple feedback mechanism that will allow beta users to score the result and references for quality based on whether or not it was relevant to the search query
5. Collect and store this feedback data

Phase 1 outcome:
A website or Chrome plugin that will be opened to select beta users.

Phase 2: Training dataset curation and model training

Work involved:
1. Clean and process data collected in phase 1.
2. Using processed data from (1) as well as previously tagged data, generate training dataset for training an open source embedding model
3. Fine tune an open source model using dataset generated in the previous phase;eplace the off-the-shelf model in the search tool with the fine-tuned model
4. Continue assessing search quality

Phase 2 outcome:
Website or chrome plugin developed in Phase 1 should have a UI for gathering feedback
Training dataset created should be open sourced in Huggingface
Open source the custom model in Huggingface

Bonus Phase: Ideate and build an innovative system to auto-detect material related to reforestation

The goal is to build an innovative way to auto-detect/crowd-source/expert-curate good quality resources for reforestation using tools built in earlier phases, and make it available to the broader community.

Implementation details for this phase are not fully thought out at this stage. This is an opportunity for creative individuals to come together to ideate and build.

Bonus Phase outcome: TBD


Coastal Wetland Forests
Elliott White Jr

The goal of our lab is to create a high-spatial resolution map of coastal forested wetlands at global scale. If we know precisely where these ecologically critical but fragile forests are located, we can manage freshwater flows to counteract saltwater introgression due to rising sea levels, and we can assist in their migration inland, preserving their critical function in protecting coastlines and sequestering carbon.

Bison
Jason Baldes
Gisel Booman

Across the continent, a number of first nations are in the process of reintroducing bison to the grasslands in which they were once the primary grazer and an ecologically vital species. Initial experiences and evolutionary considerations suggest that this may be ecologically beneficial in terms of grassland biodiversity, carbon cycle, and resilience to climate change. However, these questions have not yet been studied at scale. In this lab, we will leverage remote sensing to scale up from ground measurements, establishing the large-scale patterns of bison impact.

Riparian Ecosystems
Forrest Pound

Beaver dams are known to result in greener, more drought-resilient waterways in semi-arid environments. We are using computer vision to spot dams in satellite imagery, generating a large dataset that we can use to train models that will tell us what the ecological effects of a dam will be at any point on a waterway. The goal is to create a tool to guide efficient restoration through the introduction of small dams.

Bundled Ecological NFT
Philip Taylor

Markets in voluntary carbon credits are increasingly providing a flow of capital for regenerating ecosystems. The problem is, thriving and resilient ecosystems are not just carbon. We need to find ways to structure credits to incentivize the diverse and functional ecosystems we want, not merely high-concentrations of carbon. We will design the technological tools to support a market in bundled ecological credits.

Global Forests
Aron Boettcher

We are building an accurate and global model for predicting potential rates of reforestation and resulting carbon sequestration. Such a model could have a transformational impact on global reforestation efforts by opening new streams of financing in the form of carbon credit futures.

Impact & Risk
Aaron Hirsh
Valérie Lechêne

Leveraging The Earthshot Institute’s broad scientific and technical expertise, the Impact and Risk Lab helps investors and governments who earnestly want to forecast, measure, and address the socio-ecological risks to and/or impacts from their work. For a given system, we build simple process-based models to identify key socio-ecological risks and outcomes. We then draw on big data to improve and train our models, generating quantitative predictions and developing measurement systems for verification.

Pitch!
No items found.
Register for the

Earthshot Institute Launch Event
Fri April 22
Online