# Emotions and Courtship in Pair Bonds: Agent-Based Evolutionary Simulations of the Iterated Prisoner's Dilemma (Google DeepMind, PNAS 2023)

## Facts

- Provider: Google DeepMind (Suzanne Sadedin, Edgar A. Duenez-Guzman, Joel Z. Leibo)
- License: CC-BY-4.0 (https://github.com/google-deepmind/emotions_courtship)
- Stored at: Google Cloud Storage, gs://emotions_courtship/data (us)
- Access: public, readable without an account
- Category: Life Sciences
- Source: https://www.pnas.org/doi/abs/10.1073/pnas.2308911120
- Page: https://scigantic.com/archive/a98da2bb-6c34-56d9-9ad6-7e23f874e1a0
- Keywords: Life Sciences, Google Cloud Storage

## About

Simulation output behind an evolutionary model of how animals in coordinated pair bonds, common in birds and other taxa, solve social dilemmas with a partner. The model combines emotional bookkeeping (neuroendocrine feedback that tracks a partner's behavior) and courtship (low-stakes social dilemmas played before and during the bond), and evolves agent genotypes by agent-based simulation of the noisy iterated prisoner's dilemma. The paper finds that emotional bookkeeping and courtship sustain cooperation, especially together, but when deceitful defection is cheap, courtship often raises cooperation while emotional bookkeeping lowers it. The data/ folder has three parts: full_cross/ (full_cross.feather, about 594 MB, the main crossed parameter sweep), genotypes/ (evolved agent genotypes) and sensitivity/ (sensitivity analyses); the same data are bundled as data.zip (951 MB) at the bucket root. Read Feather files with pandas.read_feather or pyarrow; the analysis notebook is in google-deepmind/emotions_courtship. Used in behavioral ecology, evolution of cooperation and game theory. License: CC BY 4.0 for the materials (Google DeepMind). Cite Sadedin, Duenez-Guzman and Leibo, Emotions and courtship help bonded pairs cooperate, but emotional agents are vulnerable to deceit, PNAS 120(46):e2308911120, 2023 (doi 10.1073/pnas.2308911120).

## Load the data

```python
import gcsfs
fs = gcsfs.GCSFileSystem(token='anon')
fs.ls('emotions_courtship/data/', detail=False)[:10]
```

Lists the files with anonymous access. Runs on any machine, with no cloud account.

## Data structure

Example file: `data/full_cross/full_cross.feather`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/a98da2bb-6c34-56d9-9ad6-7e23f874e1a0
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/a98da2bb-6c34-56d9-9ad6-7e23f874e1a0
- Over MCP: `get_schema_card` with id `a98da2bb-6c34-56d9-9ad6-7e23f874e1a0` at https://api.scigantic.com/mcp

## Related topics

- [Life Sciences datasets](https://scigantic.com/topics/life-sciences)
