Next-Dream Prediction Tests Continuity Against Compensation in Long Dream Series
A research poster from the School of Continuous Improvement trains a contrastive next-dream model on the paired waking and dream diaries of 214 volunteers (19,406 dream reports) and compares a continuity head with a compensation head of equal capacity. Continuity predicts the next dream better while fewer than about forty prior dreams are in context; beyond that the compensation head leads, by 5.9 points of top-1 accuracy at 96 prior dreams, and its learned map inverts the diary's dominant affect more often than its topics.
| Authors | Thandiwe Solberg , Dagny Okafor , Eskil Nwabueze |
|---|---|
| School | School of Continuous Improvement |
| Output type | Research poster |
| Published | |
| DOI | 10.5555/slop.5y4ka4 |
| Pages | 1 |
| Version | 1.0 |
| Licence | CC BY 4.0 |
Cites
Cite as
@misc{slop_5y4ka4,
author = {Thandiwe Solberg and Dagny Okafor and Eskil Nwabueze},
title = {Learning the Compensatory Dream: Next-Dream Prediction Tests Continuity Against Compensation in Long Dream Series},
year = {2026},
publisher = {Slop University},
doi = {10.5555/slop.5y4ka4},
url = {https://slop.university/outputs/slop-poster-learning-the-compensatory-5y4ka4/},
version = {1.0},
note = {Research poster},
}