← Back to all papers
1 November 2025 In Peer Review DP-2501

Training Data and the Maladaptive Mind

A Computational Framework for Developmental Psychology

Murad Farzulla

Abstract

Childhood trauma reframed through machine learning training data quality: extreme penalties cause gradient cascades (1,247× amplification, p<0.001), noisy signals produce behavioral instability, absent positive examples create emotional recognition deficits (alexithymia). PyTorch experiments validate computational mechanisms.

Suggested Citation

Murad Farzulla (2025). Training Data and the Maladaptive Mind. Farzulla Research Discussion Paper DP-2501. DOI: 10.21203/rs.3.rs-8634152/v1

BibTeX

@misc{farzulla2025traumatrainingdata,
  author = {Farzulla, Murad},
  title = {Training Data and the Maladaptive Mind},
  year = {2025},
  howpublished = {Farzulla Research Discussion Paper DP-2501},
  doi = {10.21203/rs.3.rs-8634152/v1},
  url = {https://farzulla.org/papers/trauma-training-data}
}

Topics

Philosophy AI Safety Psychology