Collab is a package that supports the use of Bayesian and causal inference with Pyro and ChiRho in the study of animal collaborative behavior. The current version focuses on foraging animals and the probabilistic identification of foraging strategies. All the functionalities are illustrated in Jupyter notebooks (listed further down). Using the package you can:
- Simulate synthetic data of foraging animals (random walkers, foragers following only food trace, foragers following a leader, foragers communicating about the position of food).
- Expand a (real or simulated) dataset by calculating a set of predictor scores at each location each agent may decide to move to. Scores correspond to a relevant feature agents may value, such as presense of a food trace, proximity to other foragers, and availability of information communicated from other agents.
- Use the expanded data to profile the foraging strategy using Bayesian inference, to assess how strongly each feature drives agents' decisions of where to move.
- Compartmentalize synthetic or real-world animal movement data in preparation for Bayesian dynamical systems inference.
- Build your own dynamical systems model of the compartmentalized data and use it within a Bayesian inferential workflow.
This repository is a work in progress, with more documentation and functionality coming soon. Please reach out if you're interested in collaborating or contributing.
Basic Setup:
git clone git@github.com:BasisResearch/collaborative-intelligence.git
cd collaborative-intelligence
git checkout main
pip install .Dev Setup:
To install dev dependencies needed to contribute to Collab, run the following command:
pip install -e ".[test]"or
pip install -e .[test]Contributing:
Before submitting a pull request, please autoformat code and ensure that unit tests pass locally:
make format # runs black and isort make lint # linting make tests # notebook and unit tests
All the notebooks are located in the docs (especially docs/foraging) folder. The following notebooks are available:
- random_hungry_followers.ipynb uses of the package to simulate data of foraging animals and to use it to profile the foraging strategy (random/food trace focus/followers) using Bayesian inference.
- communicators_simulations.ipynb illustrates the use of the package to simulate data of foraging animals where the animals reveal the position of food to each other.
- communicators_inference.ipynb shows how to use our toolkit to profile this foraging strategy using Bayesian inference.
- central_park_birds_predictors.ipynb illustrates how to use the package to expand a real world dataset that includes the movement of foraging birds in Central Park, New York, into one that at each frame assigns various predictor scores to space-time points per forager.
- central_park_birds_inference.ipynb contains an example that involves using the expanded data to profile the proximity to other animals preferences of ducks and sparrows using Bayesian inference.
- locust_approximate_pipeline.ipynb goes through an analogous workflow with a real-world dataset of foraging locust, related to Information integration for decision-making in desert locusts by Günzel, Oberhauser and Couzin-Fuchs.
- locust_ds_data.ipynb illustrates how to compartmentalize the locust data in preparation for Bayesian dynamical systems inference.
- locust_ds_class.ipynb shows how to build a dynamical systems mode of the compartmentalized data and use it within the Bayesian inferential workflow.
- locust_ds_inference.ipynb shows how to build a dynamical systems model of the compartmentalized data and use it within a Bayesian inferential workflow.
- locust_ds_validate.ipynb uses the class we defined to validate the dynamical systems model of the compartmentalized data.
- locust_ds_interpret.ipynb elaborates on a proper way to interpret the inference results of the dynamical systems model.
Note: The inference steps assume some familiarity with Pyro and probabilistic programming. The Pyro repository contains links to introductory Pyro tutorials. The dynamical systems materials assume some familarity with ChiRho (see especially this tutorial).