A new paper proposes a formal framework for decision theory using nonparametric structural equation models (NPSEMs), aiming to provide a clearer modeling language and evaluative ground for the field. This approach introduces 'personal decision theory,' which guides agents to maximize their subjective counterfactual utility. The paper also suggests a performance metric for decision theories and analyzes classic problems like the smoking lesion and Newcomb's problem. Separately, a LessWrong post critiques functional decision theory (FDT), arguing it is underspecified and its recommendations are implausible, contrasting it with causal decision theory (CDT) and evidential decision theory (EDT). AI
IMPACT This research could refine AI agent decision-making frameworks, while critiques of existing theories highlight ongoing challenges in rational agent design.
RANK_REASON The cluster contains an academic paper on a theoretical framework and an opinion piece critiquing a specific decision theory.
- Ben Levinstein
- causal decision theory
- evidential decision theory
- functional decision theory
- LessWrong
- Rationalism
- Will MacAskill
- Wolfgang Schwarz
- A causal modeling perspective on decision theory
- arXiv
- causality
- decision theory
- Newcomb's problem
- nonparametric structural equation models
- NPSEMs
- personal decision theory
- philosophy
- probability
- smoking lesion problem
AI-generated summary · Google Gemini · from 3 sources. How we write summaries →
