Skip to content

Commit 7c5527f

Browse files
author
Quarto GHA Workflow Runner
committed
Built site for gh-pages
1 parent 574b095 commit 7c5527f

29 files changed

Lines changed: 349 additions & 323 deletions

‎.nojekyll‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1 +1 @@
1-
9568e30b
1+
4b2c2764

‎_tex/index.tex‎

Lines changed: 21 additions & 21 deletions
Original file line numberDiff line numberDiff line change
@@ -644,9 +644,9 @@ \subsubsection{Unit III --- Learning from Processing Data (Weeks
644644

645645
\begin{center}\rule{0.5\linewidth}{0.5pt}\end{center}
646646

647-
\subsubsection{Unit IV --- Characterization, Automation, and Uncertainty
648-
(Weeks
649-
9--11)}\label{unit-iv-characterization-automation-and-uncertainty-weeks-911}
647+
\subsubsection{Unit IV --- Characterization, Transformers, and
648+
Uncertainty (Weeks
649+
9--11)}\label{unit-iv-characterization-transformers-and-uncertainty-weeks-911}
650650

651651
\paragraph{Week 9 -- ML for characterization
652652
signals}\label{week-9-ml-for-characterization-signals}
@@ -657,9 +657,6 @@ \subsubsection{Unit IV --- Characterization, Automation, and Uncertainty
657657
\textbf{Slides:}
658658
\href{https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit09_characterization_signals/10_characterization_signals.html}{Open}
659659

660-
Companion deck (within Week 9): \textbf{Transformers for materials} ---
661-
\href{https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit09b_transformers_for_materials/transformers_for_materials.html}{Open}
662-
663660
\begin{itemize}
664661
\tightlist
665662
\item
@@ -692,43 +689,46 @@ \subsubsection{Unit IV --- Characterization, Automation, and Uncertainty
692689

693690
\begin{center}\rule{0.5\linewidth}{0.5pt}\end{center}
694691

695-
\paragraph{Week 10 -- Automation in microscopy and
696-
characterization}\label{week-10-automation-in-microscopy-and-characterization}
692+
\paragraph{Week 10 -- Transformers for materials
693+
characterization}\label{week-10-transformers-for-materials-characterization}
697694

698695
\emph{Lecture: Tuesday, 16.06.2026, 14:15-15:45 \textbar{} Exercise:
699696
Thursday, 18.06.2026, 16:15-17:45}
700697

701698
\textbf{Slides:}
702-
\href{https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit10_automation/11_automation.html}{Open}
699+
\href{https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit10_transformers_for_materials/transformers_for_materials.html}{Open}
703700

704701
\begin{itemize}
705702
\tightlist
706703
\item
707-
Autofocus, drift correction, parameter selection.
704+
Why attention: long-range correlations beyond CNN receptive fields.
705+
\item
706+
Scaled dot-product attention and the Vision Transformer (ViT).
708707
\item
709-
ML as a control component, not just a predictor.
708+
Flash Attention for tractable long sequences.
710709
\end{itemize}
711710

712711
\textbf{Summary:}
713712

714713
\begin{itemize}
715714
\tightlist
716715
\item
717-
\textbf{Autonomous characterization}: ML moves from passive analysis
718-
to active instrument control
716+
\textbf{Self-attention} and the \textbf{Vision Transformer (ViT)} for
717+
materials imaging
719718
\item
720-
\textbf{Multi-modal data fusion} (SEM + EDS + process logs) via
721-
Bayesian frameworks
719+
\textbf{Flash Attention}: long sequences without the L×L memory
720+
blow-up
722721
\item
723-
\textbf{Reinforcement learning} for instrument tuning and process
724-
optimization
722+
Applications: \textbf{ViT on 4D-STEM} diffraction; cross-attention
723+
across LPBF layer stacks
725724
\item
726-
Pipelines that autonomously find → characterize → decide the next
727-
experiment
725+
Scaling alternatives (\textbf{Mamba / state-space models}) --- and
726+
when \emph{not} to reach for a transformer
728727
\end{itemize}
729728

730-
\textbf{Exercise:} Implement a simple ML-assisted autofocus or defect
731-
detector.
729+
\textbf{Exercise:} Apply a small ViT / attention model to a
730+
characterization dataset (e.g.~4D-STEM patches); compare against a CNN
731+
baseline.
732732

733733
\begin{center}\rule{0.5\linewidth}{0.5pt}\end{center}
734734

‎index-meca.zip‎

360 Bytes
Binary file not shown.

‎index-preview.html‎

Lines changed: 14 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -256,7 +256,7 @@ <h2 id="toc-title">Table of contents</h2>
256256
<li><a href="#unit-i-experimental-data-as-a-learning-problem-weeks-13" id="toc-unit-i-experimental-data-as-a-learning-problem-weeks-13" class="nav-link" data-scroll-target="#unit-i-experimental-data-as-a-learning-problem-weeks-13"><span class="header-section-number">1.3.1</span> Unit I — Experimental Data as a Learning Problem (Weeks 1–3)</a></li>
257257
<li><a href="#unit-ii-representation-learning-for-microstructures-weeks-46" id="toc-unit-ii-representation-learning-for-microstructures-weeks-46" class="nav-link" data-scroll-target="#unit-ii-representation-learning-for-microstructures-weeks-46"><span class="header-section-number">1.3.2</span> Unit II — Representation Learning for Microstructures (Weeks 4–6)</a></li>
258258
<li><a href="#unit-iii-learning-from-processing-data-weeks-78" id="toc-unit-iii-learning-from-processing-data-weeks-78" class="nav-link" data-scroll-target="#unit-iii-learning-from-processing-data-weeks-78"><span class="header-section-number">1.3.3</span> Unit III — Learning from Processing Data (Weeks 7–8)</a></li>
259-
<li><a href="#unit-iv-characterization-automation-and-uncertainty-weeks-911" id="toc-unit-iv-characterization-automation-and-uncertainty-weeks-911" class="nav-link" data-scroll-target="#unit-iv-characterization-automation-and-uncertainty-weeks-911"><span class="header-section-number">1.3.4</span> Unit IV — Characterization, Automation, and Uncertainty (Weeks 9–11)</a></li>
259+
<li><a href="#unit-iv-characterization-transformers-and-uncertainty-weeks-911" id="toc-unit-iv-characterization-transformers-and-uncertainty-weeks-911" class="nav-link" data-scroll-target="#unit-iv-characterization-transformers-and-uncertainty-weeks-911"><span class="header-section-number">1.3.4</span> Unit IV — Characterization, Transformers, and Uncertainty (Weeks 9–11)</a></li>
260260
<li><a href="#unit-v-physics-trust-and-synthesis-weeks-1213" id="toc-unit-v-physics-trust-and-synthesis-weeks-1213" class="nav-link" data-scroll-target="#unit-v-physics-trust-and-synthesis-weeks-1213"><span class="header-section-number">1.3.5</span> Unit V — Physics, Trust, and Synthesis (Weeks 12–13)</a></li>
261261
</ul></li>
262262
<li><a href="#learning-outcomes" id="toc-learning-outcomes" class="nav-link" data-scroll-target="#learning-outcomes"><span class="header-section-number">1.4</span> Learning Outcomes</a></li>
@@ -554,13 +554,12 @@ <h4 data-number="1.3.3.2" class="anchored" data-anchor-id="week-8-inverse-proble
554554
<hr>
555555
</section>
556556
</section>
557-
<section id="unit-iv-characterization-automation-and-uncertainty-weeks-911" class="level3" data-number="1.3.4">
558-
<h3 data-number="1.3.4" class="anchored" data-anchor-id="unit-iv-characterization-automation-and-uncertainty-weeks-911"><span class="header-section-number">1.3.4</span> Unit IV — Characterization, Automation, and Uncertainty (Weeks 9–11)</h3>
557+
<section id="unit-iv-characterization-transformers-and-uncertainty-weeks-911" class="level3" data-number="1.3.4">
558+
<h3 data-number="1.3.4" class="anchored" data-anchor-id="unit-iv-characterization-transformers-and-uncertainty-weeks-911"><span class="header-section-number">1.3.4</span> Unit IV — Characterization, Transformers, and Uncertainty (Weeks 9–11)</h3>
559559
<section id="week-9-ml-for-characterization-signals" class="level4" data-number="1.3.4.1">
560560
<h4 data-number="1.3.4.1" class="anchored" data-anchor-id="week-9-ml-for-characterization-signals"><span class="header-section-number">1.3.4.1</span> Week 9 – ML for characterization signals</h4>
561561
<p><em>Lecture: Tuesday, 09.06.2026, 14:15-15:45 | Exercise: Thursday, 11.06.2026, 16:15-17:45</em></p>
562562
<p><strong>Slides:</strong> <a href="https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit09_characterization_signals/10_characterization_signals.html">Open</a></p>
563-
<p>Companion deck (within Week 9): <strong>Transformers for materials</strong> — <a href="https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit09b_transformers_for_materials/transformers_for_materials.html">Open</a></p>
564563
<ul>
565564
<li>Spectral data: XRD, EELS, EDS.</li>
566565
<li>Denoising, peak finding, dimensionality reduction.</li>
@@ -576,22 +575,23 @@ <h4 data-number="1.3.4.1" class="anchored" data-anchor-id="week-9-ml-for-charact
576575
<p><strong>Exercise:</strong> Apply PCA/NMF to spectral datasets; interpret components physically.</p>
577576
<hr>
578577
</section>
579-
<section id="week-10-automation-in-microscopy-and-characterization" class="level4" data-number="1.3.4.2">
580-
<h4 data-number="1.3.4.2" class="anchored" data-anchor-id="week-10-automation-in-microscopy-and-characterization"><span class="header-section-number">1.3.4.2</span> Week 10 – Automation in microscopy and characterization</h4>
578+
<section id="week-10-transformers-for-materials-characterization" class="level4" data-number="1.3.4.2">
579+
<h4 data-number="1.3.4.2" class="anchored" data-anchor-id="week-10-transformers-for-materials-characterization"><span class="header-section-number">1.3.4.2</span> Week 10 – Transformers for materials characterization</h4>
581580
<p><em>Lecture: Tuesday, 16.06.2026, 14:15-15:45 | Exercise: Thursday, 18.06.2026, 16:15-17:45</em></p>
582-
<p><strong>Slides:</strong> <a href="https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit10_automation/11_automation.html">Open</a></p>
581+
<p><strong>Slides:</strong> <a href="https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit10_transformers_for_materials/transformers_for_materials.html">Open</a></p>
583582
<ul>
584-
<li>Autofocus, drift correction, parameter selection.</li>
585-
<li>ML as a control component, not just a predictor.</li>
583+
<li>Why attention: long-range correlations beyond CNN receptive fields.</li>
584+
<li>Scaled dot-product attention and the Vision Transformer (ViT).</li>
585+
<li>Flash Attention for tractable long sequences.</li>
586586
</ul>
587587
<p><strong>Summary:</strong></p>
588588
<ul>
589-
<li><strong>Autonomous characterization</strong>: ML moves from passive analysis to active instrument control</li>
590-
<li><strong>Multi-modal data fusion</strong> (SEM + EDS + process logs) via Bayesian frameworks</li>
591-
<li><strong>Reinforcement learning</strong> for instrument tuning and process optimization</li>
592-
<li>Pipelines that autonomously find → characterize → decide the next experiment</li>
589+
<li><strong>Self-attention</strong> and the <strong>Vision Transformer (ViT)</strong> for materials imaging</li>
590+
<li><strong>Flash Attention</strong>: long sequences without the L×L memory blow-up</li>
591+
<li>Applications: <strong>ViT on 4D-STEM</strong> diffraction; cross-attention across LPBF layer stacks</li>
592+
<li>Scaling alternatives (<strong>Mamba / state-space models</strong>) — and when <em>not</em> to reach for a transformer</li>
593593
</ul>
594-
<p><strong>Exercise:</strong> Implement a simple ML-assisted autofocus or defect detector.</p>
594+
<p><strong>Exercise:</strong> Apply a small ViT / attention model to a characterization dataset (e.g.&nbsp;4D-STEM patches); compare against a CNN baseline.</p>
595595
<hr>
596596
</section>
597597
<section id="week-11-uncertainty-aware-regression-gaussian-processes" class="level4" data-number="1.3.4.3">

‎index.docx‎

22 Bytes
Binary file not shown.

‎index.embed.ipynb‎

Lines changed: 15 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,7 @@
1313
"\n",
1414
"This course teaches how machine learning can be applied to experimental data from materials processing and characterization. The focus lies on images, spectra, time-series, and processing parameters, and on understanding how physical data formation interacts with learning algorithms. Students learn to build robust, uncertainty-aware ML pipelines for real experimental workflows, avoiding common pitfalls such as data leakage, overfitting, and spurious correlations."
1515
],
16-
"id": "67ba1aeb-03cd-4d92-a5f7-46e3302efa0f"
16+
"id": "8e3b1a1b-18ab-4c3d-90b6-70426be6db43"
1717
},
1818
{
1919
"cell_type": "raw",
@@ -78,7 +78,7 @@
7878
"}\n",
7979
"</style>"
8080
],
81-
"id": "289420e6-fce3-4e34-a488-3a338e0de8d0"
81+
"id": "598bd7dc-64ed-4353-9ac5-4df6a42f8f71"
8282
},
8383
{
8484
"cell_type": "raw",
@@ -108,7 +108,7 @@
108108
" <strong>How to use this course site.</strong> Use this page as the central hub for syllabus, lecture structure, reading, notebooks, and course materials. Formal announcements and enrollment remain on StudOn; code and openly shared resources live in the linked GitHub repository.\n",
109109
"</div>"
110110
],
111-
"id": "18eff77b-b70f-44bb-8412-4ff5b77d375b"
111+
"id": "1c001372-5073-4a25-b13d-2c8ad86690c9"
112112
},
113113
{
114114
"cell_type": "markdown",
@@ -332,16 +332,14 @@
332332
"\n",
333333
"------------------------------------------------------------------------\n",
334334
"\n",
335-
"### Unit IV — Characterization, Automation, and Uncertainty (Weeks 9–11)\n",
335+
"### Unit IV — Characterization, Transformers, and Uncertainty (Weeks 9–11)\n",
336336
"\n",
337337
"#### Week 9 – ML for characterization signals\n",
338338
"\n",
339339
"*Lecture: Tuesday, 09.06.2026, 14:15-15:45 \\| Exercise: Thursday, 11.06.2026, 16:15-17:45*\n",
340340
"\n",
341341
"**Slides:** [Open](https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit09_characterization_signals/10_characterization_signals.html)\n",
342342
"\n",
343-
"Companion deck (within Week 9): **Transformers for materials** — [Open](https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit09b_transformers_for_materials/transformers_for_materials.html)\n",
344-
"\n",
345343
"- Spectral data: XRD, EELS, EDS.\n",
346344
"- Denoising, peak finding, dimensionality reduction.\n",
347345
"- Using ML without destroying physical meaning.\n",
@@ -357,23 +355,24 @@
357355
"\n",
358356
"------------------------------------------------------------------------\n",
359357
"\n",
360-
"#### Week 10 – Automation in microscopy and characterization\n",
358+
"#### Week 10 – Transformers for materials characterization\n",
361359
"\n",
362360
"*Lecture: Tuesday, 16.06.2026, 14:15-15:45 \\| Exercise: Thursday, 18.06.2026, 16:15-17:45*\n",
363361
"\n",
364-
"**Slides:** [Open](https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit10_automation/11_automation.html)\n",
362+
"**Slides:** [Open](https://pelzlab.science/public_presentations/ml_for_characterization_and_processing/unit10_transformers_for_materials/transformers_for_materials.html)\n",
365363
"\n",
366-
"- Autofocus, drift correction, parameter selection.\n",
367-
"- ML as a control component, not just a predictor.\n",
364+
"- Why attention: long-range correlations beyond CNN receptive fields.\n",
365+
"- Scaled dot-product attention and the Vision Transformer (ViT).\n",
366+
"- Flash Attention for tractable long sequences.\n",
368367
"\n",
369368
"**Summary:**\n",
370369
"\n",
371-
"- **Autonomous characterization**: ML moves from passive analysis to active instrument control\n",
372-
"- **Multi-modal data fusion** (SEM + EDS + process logs) via Bayesian frameworks\n",
373-
"- **Reinforcement learning** for instrument tuning and process optimization\n",
374-
"- Pipelines that autonomously find → characterize → decide the next experiment\n",
370+
"- **Self-attention** and the **Vision Transformer (ViT)** for materials imaging\n",
371+
"- **Flash Attention**: long sequences without the L×L memory blow-up\n",
372+
"- Applications: **ViT on 4D-STEM** diffraction; cross-attention across LPBF layer stacks\n",
373+
"- Scaling alternatives (**Mamba / state-space models**) — and when *not* to reach for a transformer\n",
375374
"\n",
376-
"**Exercise:** Implement a simple ML-assisted autofocus or defect detector.\n",
375+
"**Exercise:** Apply a small ViT / attention model to a characterization dataset (e.g. 4D-STEM patches); compare against a CNN baseline.\n",
377376
"\n",
378377
"------------------------------------------------------------------------\n",
379378
"\n",
@@ -457,7 +456,7 @@
457456
"- ML-assisted autofocus or EBSD pattern classification.\n",
458457
"- Multi-modal fusion of images, spectra, and process parameters."
459458
],
460-
"id": "49e3c9e9-5a28-4836-a54e-1a1c3f84a40a"
459+
"id": "21b74023-ae88-4b57-95eb-8888923c21ba"
461460
}
462461
],
463462
"nbformat": 4,

0 commit comments

Comments
 (0)