UH Biocomputation Group - Stavros Anagnouhttp://biocomputation.herts.ac.uk/2022-02-16T10:15:52+00:00Spurious Normativity Enhances Learning of Compliance and Enforcement Behavior in Artificial Agents2022-02-16T10:15:52+00:002022-02-16T10:15:52+00:00Stavros Anagnoutag:biocomputation.herts.ac.uk,2022-02-16:/2022/02/16/spurious-normativity-enhances-learning-of-compliance-and-enforcement-behavior-in-artificial-agents.html<p class="first last">Stavros Anagnou's Journal Club session where he will talk about a paper "Spurious Normativity Enhances Learning of Compliance and Enforcement Behavior in Artificial Agents"</p>
<p>This week on Journal Club session Stavros Anagnou will talk about a paper
"Spurious Normativity Enhances Learning of Compliance and Enforcement Behavior
in Artificial Agents".</p>
<hr class="docutils" />
<p>How do societies learn and maintain social norms? Here we use multiagent
reinforcement learning to investigate the learning dynamics of enforcement and
compliance behaviors. Artificial agents populate a foraging environment and
need to learn to avoid a poisonous berry. Agents learn to avoid eating
poisonous berries better when doing so is taboo, meaning the behavior is
punished by other agents. The taboo helps overcome a credit assignment problem
in discovering delayed health effects. Critically, introducing an additional
taboo, which results in punishment for eating a harmless berry, further
improves overall returns. This "silly rule" counterintuitively has a positive
effect because it gives agents more practice in learning rule enforcement. By
probing what individual agents have learned, we demonstrate that normative
behavior relies on a sequence of learned skills. Learning rule compliance
builds upon prior learning of rule enforcement by other agents. Our results
highlight the benefit of employing a multiagent reinforcement learning
computational model focused on learning to implement complex actions.</p>
<div class="line-block">
<div class="line"><br /></div>
</div>
<p>Papers:</p>
<ul class="simple">
<li>R. Köster, D. Hadfield-Menell, R. Everett, L. Weidinger, G. Hadfield, J. Leibo, <a class="reference external" href="https://doi.org/10.1073/pnas.2106028118">"Spurious Normativity Enhances Learning of Compliance and Enforcement Behavior in Artificial Agents"</a>, 2022, Proceedings of the National Academy of Sciences, 119, e2106028118</li>
</ul>
<p><strong>Date:</strong> 2022/02/18 <br />
<strong>Time:</strong> 14:00 <br />
<strong>Location</strong>: online</p>