Shark AI is a collaborative AI research project focused on integrating machine learning with paleontology, specifically in the study of fossil shark teeth. The project is part of a larger NSF-funded initiative (Grant No. 2147625) aimed at enhancing STEM education in Florida's public middle schools by incorporating fossil research into the curriculum. The AI component of the project, referred to as PaleoAI, is a spin-off from the primary educational research objectives and aims to address basic research questions in paleontology using advanced machine learning techniques.
- NSF Funding: This project is funded by the National Science Foundation (NSF) under Grant No. 2147625 to the University of Florida (UF).
- Educational Impact: The research integrates fossil shark teeth studies into Florida public middle schools, involving 50 teachers, over 10,000 students, across 30 counties.
- Research Team: The research is conducted by a team of paleontologists, biologists, and software engineers, including professionals and volunteers from UF and other institutions.
- PaleoAI: A focused effort on applying machine learning to paleontological data, specifically fossil shark teeth, to answer research questions about species identification and classification.
The Docker images used in this project are fully compatible with NVIDIA A100 GPUs, taking full advantage of the GPU's Tensor Core architecture for accelerated deep learning workloads.
- Download Docker Desktop for Mac from Docker Hub.
- Follow the installer's instructions to complete the installation.
- Clone the Repository
Clone the project repository to your local machine using:
git clone https://github.com/brberis/djtensor.git
- Environment Configuration
Copy the .env.dev.example file to create a .env.dev file and update the environment variables as needed:
- Build the Containers
Navigate to the project directory where the docker-compose.yml file is located and run the following command to build the Docker containers:
docker-compose build
- Start the Services
Start the services defined in the Docker Compose configuration file with:
docker-compose up
Alternatively, you can run them in detached mode using:
docker-compose up -d
- Access the Application
Once the containers are up and running, you can access the frontend application by navigating to http://localhost:3000 in your browser. The backend API will be available at http://localhost:8000. Username: testuser Password: testpassword.
To stop all services, you can run:
docker-compose down
- Bruce MacFadden - Distinguished Professor and Curator of Vertebrate Paleontology at the Florida Museum of Natural History, and Founding Director of the Thompson Earth Systems Institute. With a career at UF since 1977, Bruce has authored several seminal books and over 200 peer-reviewed articles. He has secured more than $35 million in research funding, primarily from NSF, and has held prestigious positions, including President of both the Paleontological Society and the Society of Vertebrate Paleontology. His research spans fossil mammals, paleobiology, and macroevolution, with significant contributions to understanding the biodiversity of the New World tropics through work in the Panama Canal region. Bruce is also heavily involved in K-12 STEM education outreach and is the principal investigator for the current NSF project integrating AI and paleontology in middle schools:contentReference[oaicite:0]{index=0}.
- Victor Perez - Co-Principal Investigator (Co-PI), PhD from UF, specializing in paleontology.
- Arthur Porto - Curator of Artificial Intelligence for Natural History and Biodiversity at the Florida Museum of Natural History. Arthur Porto has a unique background in both biology and machine learning, specializing in using AI to study evolutionary patterns. Prior to his current role, he was an assistant professor at Louisiana State University, where he focused on developing tools like ALPACA to automate the analysis of 3D models of biological specimens:contentReference[oaicite:0]{index=0}.
- Ken Marks - An experienced researcher with a focus on vertebrate paleontology and fossil shark teeth. Ken Marks has contributed extensively to the study of shark fossils, particularly in Florida, and has been involved in numerous projects aimed at understanding the evolutionary history of these ancient creatures.
- Stephanie Killingsworth - PhD student in Geology at UF, focusing on fossil research and paleontology.
- Maria Camila Vallejos - PhD candidate in Biology at UF, contributing to the integration of biological sciences with paleontological research.
- Samantha Zbinden - UF graduate, PhD student at UT-Austin starting August 2024, involved in fossil shark research.
-
Cristobal Barberis
Lead developer responsible for the entire software architecture, implementation, and deployment of the project. Led the development of the machine learning models, backend, and integration processes. -
Arthur Porto
AI expert and advisor who provided domain-specific insights into machine learning applications for paleontology. Contributed by reviewing models and writing code for key AI functionalities. -
Dêvi Hall
Contributor who assisted in project coordination, documentation, and process workflows, ensuring alignment between research and software development.
This software is licensed under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0).
- Academic Use: Free for research, teaching, and educational purposes.
- Commercial Use: Prohibited without prior written permission from the author.
- For commercial inquiries, please contact Cristobal Barberis at [email protected].