Add Computational Mathematics Research Digest

README.md

@@ -10,7 +10,8 @@ same six-page structure (Home, Publications, People, News, Projects, Contact).
 index.html          Home — bio, positions, education, service
 publications.html   All 38 publications (2015–2026), from your CV
 people.html         Students supervised, examinerships, collaborators
-news.html           Empty for now — see comment in the file for how to add items
+news.html           Index of the Computational Mathematics Research Digest
+news/               Permanent pages for individual weekly digest issues
 projects.html       CodeLaTeX, claude-latex-skill, outreach/video content
 contact.html        Emails, offices, ORCID/Scopus/WoS/Scholar, social links
 style.css           All styling (one shared stylesheet)
@@ -33,9 +34,8 @@ Everything on the site is now sourced from your CV (March 2026) or from things y
   button. If that changes later, just ask and I'll build a public-safe version.
 
 Smaller open items:
-- **news.html** is intentionally empty (you said no news for now). There's a commented template
-  in the file itself for when you're ready to add the first item — natural first entries would be
-  the October 2025 promotion/appointment or the January 2026 Sultan Qaboos University talk.
+- **news.html** now indexes the weekly Computational Mathematics Research Digest. Each issue has a
+  permanent, shareable page under `news/`, with newest issues displayed first.
 - **projects.html** — CodeLaTeX is marked "in development" with no public link, per your note that
   it's still under construction. Ask and I'll add a repo/demo link once it's ready to share.
 
@@ -75,8 +75,12 @@ authors/title/venue/link, and renumber every `<span class="pub-num">[N]</span>`
 paper = highest number, shown first — the `reversed` attribute on the `<ol>` just controls the visual
 counting direction, so the numbers themselves still need to be right).
 
-**New student / news item / project** — copy the relevant block (`.person`, `.news-item`, or
-`.project`) in the corresponding page and edit the text. No other page needs to change.
+**New digest issue** — copy `news/digest-template.html`, rename it using
+`YYYY-MM-DD-computational-mathematics-digest.html`, replace the placeholders, and add a new
+`.digest-card` at the top of `news.html`. Update the previous issue's “Newer issue” navigation link.
+
+**New student / project** — copy the relevant `.person` or `.project` block in the corresponding
+page and edit the text. No other page needs to change.
 
 ## Custom domain (optional)
 

news.html

@@ -4,6 +4,7 @@
 <meta charset="UTF-8">
 <meta name="viewport" content="width=device-width, initial-scale=1.0">
 <title>Hassan Mohammad — News</title>
+<meta name="description" content="Computational Mathematics Research Digest — a weekly critical selection of papers in numerical optimization, numerical linear algebra, nonlinear least squares, and nonlinear equations.">
 <link rel="preconnect" href="https://fonts.googleapis.com">
 <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
 <link href="https://fonts.googleapis.com/css2?family=Source+Serif+4:ital,wght@0,400;0,600;0,700;1,400&family=IBM+Plex+Sans:wght@400;500;600&family=IBM+Plex+Mono:wght@400;500&display=swap" rel="stylesheet">
@@ -29,20 +30,43 @@
 </header>
 
 <main>
-  <p class="eyebrow">Updates</p>
-  <h2 class="section-title">News</h2>
+  <p class="eyebrow">Weekly research briefing</p>
+  <h2 class="section-title">Computational Mathematics Research Digest</h2>
+  <p class="digest-intro">A critical weekly selection of worthwhile papers, preprints, and technical reports in numerical optimization, numerical linear algebra, nonlinear least squares, and nonlinear equations. Each issue explains why the work matters and where to begin reading.</p>
 
-  <p style="color: var(--text-muted);">Nothing posted yet — check back soon.</p>
+  <section class="digest-list" aria-label="Research digest issues">
+    <article class="digest-card">
+      <div class="digest-meta"><time datetime="2026-09-04">4 September 2026</time> · Issue 4</div>
+      <h3><a href="news/2026-09-04-computational-mathematics-digest.html">Trust regions without function values, gradient-norm complexity, and higher-order Newton methods</a></h3>
+      <p>Four selected papers, led by a decentralized objective-function-free trust-region method with centralized-order complexity guarantees.</p>
+      <div class="topic-tags"><span>Optimization</span><span>Newton methods</span><span>PDE-constrained optimization</span></div>
+      <a class="digest-link" href="news/2026-09-04-computational-mathematics-digest.html">Read issue 4 →</a>
+    </article>
 
-  <!--
-    To add an item, copy this block just above the closing </main> tag (or above this comment)
-    and edit the date and text. Newest entries go at the top.
+    <article class="digest-card">
+      <div class="digest-meta"><time datetime="2026-08-28">28 August 2026</time> · Issue 3</div>
+      <h3><a href="news/2026-08-28-computational-mathematics-digest.html">A DFP counterexample, primal Newton acceleration, and block Toeplitz conditioning</a></h3>
+      <p>Five substantive papers, including a negative resolution of a long-standing global-convergence question for classical DFP.</p>
+      <div class="topic-tags"><span>Quasi-Newton</span><span>Proximal methods</span><span>Numerical linear algebra</span><span>Nonlinear equations</span></div>
+      <a class="digest-link" href="news/2026-08-28-computational-mathematics-digest.html">Read issue 3 →</a>
+    </article>
 
-    <div class="news-item">
-      <div class="news-date">2026 · Jan</div>
-      <div>Short description of the update, in one or two sentences.</div>
-    </div>
-  -->
+    <article class="digest-card">
+      <div class="digest-meta"><time datetime="2026-08-21">21 August 2026</time> · Issue 2</div>
+      <h3><a href="news/2026-08-21-computational-mathematics-digest.html">Random-subspace DFO, single-loop min–max optimization, and matrix-free NMPC</a></h3>
+      <p>Four core selections and one strategic read on optimization-driven progress in matrix multiplication.</p>
+      <div class="topic-tags"><span>Derivative-free optimization</span><span>Min–max</span><span>Least squares</span><span>Matrix-free methods</span></div>
+      <a class="digest-link" href="news/2026-08-21-computational-mathematics-digest.html">Read issue 2 →</a>
+    </article>
+
+    <article class="digest-card">
+      <div class="digest-meta"><time datetime="2026-08-14">14 August 2026</time> · Issue 1</div>
+      <h3><a href="news/2026-08-14-computational-mathematics-digest.html">Lanczos near-optimality, dynamic proximal points, and alternating Levenberg–Marquardt</a></h3>
+      <p>The inaugural shortlist: five methodologically substantial contributions across Krylov methods, optimization, nonlinear least squares, and structured KKT systems.</p>
+      <div class="topic-tags"><span>Lanczos</span><span>Proximal point</span><span>Levenberg–Marquardt</span><span>Riccati recursion</span></div>
+      <a class="digest-link" href="news/2026-08-14-computational-mathematics-digest.html">Read issue 1 →</a>
+    </article>
+  </section>
 
 </main>
 

news/2026-08-14-computational-mathematics-digest.html

@@ -0,0 +1,9 @@
+<!DOCTYPE html><html lang="en"><head><meta charset="UTF-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><title>Computational Mathematics Research Digest — 14 August 2026</title><meta name="description" content="Research digest: Lanczos near-optimality, dynamic proximal points, alternating Levenberg–Marquardt, parallel Riccati recursions, and equilibrium QP reformulations."><link rel="preconnect" href="https://fonts.googleapis.com"><link rel="preconnect" href="https://fonts.gstatic.com" crossorigin><link href="https://fonts.googleapis.com/css2?family=Source+Serif+4:ital,wght@0,400;0,600;0,700;1,400&family=IBM+Plex+Sans:wght@400;500;600&family=IBM+Plex+Mono:wght@400;500&display=swap" rel="stylesheet"><link rel="stylesheet" href="../style.css"></head><body>
+<header class="site-header"><div class="header-inner"><p class="site-name"><a href="../index.html">Hassan Mohammad</a></p><nav class="site-nav" aria-label="Primary"><ul><li><a href="../index.html">Home</a></li><li><a href="../publications.html">Publications</a></li><li><a href="../people.html">People</a></li><li><a href="../news.html" class="active">News</a></li><li><a href="../projects.html">Projects</a></li><li><a href="../artifacts.html">Artifacts</a></li><li><a href="../contact.html">Contact</a></li></ul></nav></div></header>
+<main class="digest-article"><p class="eyebrow">Computational Mathematics Research Digest · Issue 1</p><h1>Weekly computational mathematics digest</h1><div class="article-meta">Published <time datetime="2026-08-14">14 August 2026</time> · Curated by Hassan Mohammad</div><p class="coverage">Coverage: substantive postings from 2–12 August 2026, ranked by relevance and potential research value.</p>
+<article class="paper-entry"><h2>1. Optimal Near-Optimality Bounds for the Lanczos Method for Matrix Functions</h2><p class="paper-byline"><strong>Tyler Chen and David Persson</strong> · arXiv · 7 August 2026 · <a href="https://arxiv.org/abs/2608.07160" target="_blank" rel="noopener">Primary source ↗</a></p><p><span class="paper-label">Why it matters</span><br>For Hermitian positive-definite A, the authors show that Lanczos approximations to f(A)b are within an explicit condition-number-dependent factor of the best approximation in the same Krylov space. The result covers Stieltjes-related functions and the constant is proved optimal.</p><p><span class="paper-label">Read first</span><br>Section 1.2, especially Theorems 1–2 and Figure 2, followed by Remark 3 for the sharpened Euclidean-norm interpretation for conjugate gradients.</p></article>
+<article class="paper-entry"><h2>2. Dynamic Proximal Point Method for Unconstrained Minimization</h2><p class="paper-byline"><strong>Enrico Bertolazzi, Alberto De Marchi and Davide Stocco</strong> · arXiv · 4 August 2026 · <a href="https://arxiv.org/abs/2608.03349" target="_blank" rel="noopener">Primary source ↗</a></p><p><span class="paper-label">Why it matters</span><br>The method combines adaptively scaled proximal subproblems with an inner Newton–line-search solver, including a reduced Newton system, merit function, diagonal scaling, stopping rules, and implementation-oriented pseudocode.</p><p><span class="paper-label">Read first</span><br>Algorithms 1–2 and Sections 2.0.1 and 3.1 for the outer regularization and inner solver, then Section 4’s tests on 100 benchmark problems.</p></article>
+<article class="paper-entry"><h2>3. Alternating Levenberg–Marquardt Training of Physics-Informed Neural Networks with Fourier-Enhanced Features</h2><p class="paper-byline"><strong>Yulun Wu, Matthieu Barreau, Miguel Aguiar and Karl H. Johansson</strong> · arXiv · 6 August 2026 · <a href="https://arxiv.org/abs/2608.05892" target="_blank" rel="noopener">Primary source ↗</a></p><p><span class="paper-label">Why it matters</span><br>The framework separates basis learning from coefficient fitting, handling the latter through Levenberg–Marquardt and exposing more exploitable least-squares structure than ordinary end-to-end PINN training. It includes convergence results and reports errors up to two orders of magnitude below selected baselines.</p><p><span class="paper-label">Read first</span><br>Sections 3.3–3.4 and Theorem 1 for the alternating LM framework and convergence conditions, followed by Sections 4.3–4.5 on the high-frequency heat, cavity-flow, and Burgers problems.</p></article>
+<article class="paper-entry"><h2>4. Rake–Compress Riccati Recursions for Parallel Scenario-Tree Model Predictive Control</h2><p class="paper-byline"><strong>João Sousa-Pinto</strong> · arXiv · 2 August 2026 · <a href="https://arxiv.org/abs/2608.01332" target="_blank" rel="noopener">Primary source ↗</a></p><p><span class="paper-label">Why it matters</span><br>The exact tree-contraction solver has O(N) work and storage but O(log N) parallel span, independently of tree height or balance. It reconstructs Riccati coefficients, primal variables, and multipliers, and is accompanied by JAX implementations.</p><p><span class="paper-label">Read first</span><br>Sections IV–V for the contraction construction, Theorem 3 for equivalence with the KKT system, and Section VII for parallel complexity.</p></article>
+<article class="paper-entry"><h2>5. Solving Monotone Linear-Quadratic Generalized Nash Equilibrium Problems via Quadratic Programming</h2><p class="paper-byline"><strong>Alberto Bemporad and Tatiana Tatarenko</strong> · arXiv · 7 August 2026 · <a href="https://arxiv.org/abs/2608.07336" target="_blank" rel="noopener">Primary source ↗</a></p><p><span class="paper-label">Why it matters</span><br>The principal structural result converts a monotone linear-quadratic variational GNE problem into one convex quadratic program. Regularized and accelerated variants obtain O(1/k²) approximate-equilibrium convergence, while an invertible game Jacobian permits a smaller dual-space QP.</p><p><span class="paper-label">Read first</span><br>The QP-equivalence theorem and lower-dimensional reduction, followed by the game-theoretic MPC experiment and comparisons with extragradient-type methods.</p></article>
+<aside class="category-note"><strong>Category note.</strong> No sufficiently strong new general-purpose algorithm for systems of nonlinear equations appeared; recent postings were predominantly application-specific PDE discretizations.</aside><p class="editorial-note">Selection and commentary are editorial. Dates and bibliographic details link to the cited primary sources.</p><nav class="digest-nav" aria-label="Digest issues"><span aria-current="page">First issue</span><a href="../news.html">All issues</a><a href="2026-08-21-computational-mathematics-digest.html">Issue 2 →</a></nav></main><footer class="site-footer">© 2026 Hassan Mohammad. Built with plain HTML &amp; CSS, hosted on GitHub Pages.</footer></body></html>