Learning the Value Systems of Agents with Preference-based and Inverse Reinforcement Learning
Andrés Holgado-Sánchez (ORCID: 0000-0001-8853-1022), Holger Billhardt (ORCID: 0000-0001-8298-4178), Alberto Fernández (ORCID: 0000-0002-8962-6856), Sascha Ossowski (ORCID: 0000-0003-2483-9508) · 2026
Read Learning the Value Systems of Agents with Preference-based and Inverse Reinforcement Learning by Andrés Holgado-Sánchez (ORCID: 0000-0001-8853-1022), Holger Billhardt (ORCID: 0000-0001-8298-4178), Alberto Fernández (ORCID: 0000-0002-8962-6856), Sascha Ossowski (ORCID: 0000-0003-2483-9508) (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-04518.txt.
Description
Andrés Holgado-Sánchez and 3 co-authors study how software agents could learn what their users value, in a 2026 paper on agreement technologies. These are open systems where autonomous agents negotiate on behalf of people and must reach acceptable deals.
Different people weigh moral values differently, and stating a value's exact meaning in code is hard. The authors define the value system learning problem formally. They build methods based on preference learning and inverse reinforcement learning, which infers goals from observed behavior. Their setting uses multi-objective decision processes, where several goals compete. Two simulated cases illustrate and test the approach. The goal is agreements that stay aligned with human values.
Catalog record
| CREATOR | Andrés Holgado-Sánchez (ORCID: 0000-0001-8853-1022), Holger Billhardt (ORCID: 0000-0001-8298-4178), Alberto Fernández (ORCID: 0000-0002-8962-6856), Sascha Ossowski (ORCID: 0000-0003-2483-9508) |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 113 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-04518