Program Equilibrium in the Prisoner ’ s Dilemma via Löb ’ s Theorem

By Patrick LaVictoire, Benja Fallenstein, Eliezer Yudkowsky, Mihály Bárász. 2014

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Tom Everitt
2. Marcus Hutter
3. Gary Lea
4. Andrew Critch
5. Nick Bostrom
6. Jan Leike
7. J. Urgen Schmidhuber
8. Stuart Armstrong

Citations in the corpus

• Parametric Bounded Löb's Theorem and Robust Cooperation of Bounded Agents (2016, background): “Thus, it will hopefully be of interest to game theorists and logicians alike.”
• Towards Safe Artificial General Intelligence (2018, background): “For example, when playing a prisoner’s dilemma against your own source code (Lavictoire et al., 2014), your opponent defects if and only if you defect.”
• Strategic Implications of Openness in AI Development (2017, background): “…arti fi cial agents that are capable of highly sophisticated planning and strategic reasoning (and which might be able to coordinate using different or more effective means than humans (See e.g. LaVictoire et al., 2014)), there may be additional and novel ways for systemic failures to occur.”
• AGI Safety Literature Review (2018, background): “Exceptions exist, especially in game-theoretic situations where the actions of other agents may depend on the agent’s goals or other properties (LaVictoire et al., 2014).”