Fabian Pedregosa<p>Next week I&#39;ll be at <a href="https://sigmoid.social/tags/NeurIPS2022" class="mention hashtag" rel="tag">#<span>NeurIPS2022</span></a> presenting a couple of papers. The first one is on <a href="https://sigmoid.social/tags/autodiff" class="mention hashtag" rel="tag">#<span>autodiff</span></a> through <a href="https://sigmoid.social/tags/optimization" class="mention hashtag" rel="tag">#<span>optimization</span></a> (aka <a href="https://sigmoid.social/tags/unrolling" class="mention hashtag" rel="tag">#<span>unrolling</span></a>) and its bizarre convergence properties. A 🧵 on the paper (<a href="https://arxiv.org/pdf/2209.13271.pdf" target="_blank" rel="nofollow noopener" translate="no"><span class="invisible">https://</span><span class="">arxiv.org/pdf/2209.13271.pdf</span><span class="invisible"></span></a>) (1/9)</p>