https://maddiecolbert.com/post/Recent content in Posts on Madeline Colbert's WebsiteHugoen-usTue, 10 Mar 2026 11:36:11 -0500https://maddiecolbert.com/post/iosc/Tue, 10 Mar 2026 11:36:11 -0500https://maddiecolbert.com/post/iosc/<h1 id="linear-regions">Linear Regions</h1> <p>I talked a fair bit about this idea of a binary vector that results from embedding a ReLU activated neural network into a MIP. This is denoted as $\mathscr{Z}(x)$ throughout the presentation. But where does that come from?</p> <p>Since the ReLU function is just the max between 0 and its input, we can model this using big-M constraints, $out \ge 0, out \ge in, out \le in + Mz, out \le M(1-z)$.</p>https://maddiecolbert.com/post/linear_regions/Mon, 09 Mar 2026 14:27:32 -0500https://maddiecolbert.com/post/linear_regions/<p>The following is the general MIP formulation of a ReLU function $$ out \ge in, out \ge 0$$</p>https://maddiecolbert.com/post/first_post/Mon, 09 Mar 2026 11:26:52 -0500https://maddiecolbert.com/post/first_post/<h1 id="this-is-a-first-post">This is a first post</h1> <p>This is my first post using hugo</p>