1: \begin{abstract}

2:   Isotonic regression (IR) is a non-parametric calibration method used in supervised learning.  For performing large-scale IR, we propose a primal-dual active-set (PDAS) algorithm which, in contrast to the state-of-the-art Pool Adjacent Violators (PAV) algorithm, is easily warm-started and can be parallelized.  This warm-starting capability makes it well-suited for online settings.  We prove that, like the PAV algorithm, our PDAS algorithm for IR is convergent and has a work complexity of $\mathcal{O}(n)$, though our numerical experiments suggest that our PDAS algorithm is often faster than PAV.  In addition, we propose PDAS variants (with safeguarding to ensure convergence) for solving related trend filtering (TF) problems, providing the results of experiments to illustrate their effectiveness.

3: \end{abstract}

4: