Robert Peharz<p>I&#39;m happy to share that our paper &quot;Bayesian Structure Scores for Probabilistic Circuits&quot; has been accepted at <a href="https://sigmoid.social/tags/aistats23" class="mention hashtag" rel="tag">#<span>aistats23</span></a>! This is joint work with Yang Yang (MSc @mastodon.world, incoming PhD <span class="h-card" translate="no"><a href="https://sigmoid.social/@leuvenai" class="u-url mention">@<span>leuvenai</span></a></span> and Gennaro Gala.</p><p>The main contribution of the paper is to propose a new (actually old) way to learn the structure of <a href="https://sigmoid.social/tags/ProbabilisticCircuits" class="mention hashtag" rel="tag">#<span>ProbabilisticCircuits</span></a>. We take lessons from classical structure learning in Bayesian networks which optimizes some structure score, where a principled choice is a Bayesian score.</p>