Push dev branch build

Changelog modified: true
Contributors modified: false

docs/contents/core/ai_for_good/ai_for_good.html

@@ -1120,10 +1120,10 @@ <h4 class="anchored" data-anchor-id="pattern-structure-1">Pattern Structure</h4>
 </div>
 <p>Each layer in the progressive enhancement pattern operates independently, so that systems remain functional regardless of the availability of higher tiers. The pattern’s modular structure enables seamless transitions between layers, minimizing disruptions as systems dynamically adjust to changing resource conditions. By prioritizing adaptability, the progressive enhancement pattern supports a wide range of deployment environments, from remote, resource-constrained regions to well-connected urban centers.</p>
 <p><a href="#fig-pattern-pep" class="quarto-xref">Figure&nbsp;<span>19.4</span></a> illustrates these three layers, showing the functionalities at each layer. The diagram visually demonstrates how each layer scales up based on available resources and how the system can fallback to lower layers when resource constraints occur.</p>
-<div id="fig-pattern-pep" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-env="figure" data-fig-pos="htb">
+<div id="fig-pattern-pep" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-pos="htb" data-fig-env="figure">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-pattern-pep-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<div class="sourceCode" id="fig-pattern-pep" data-fig-env="figure" data-fig-pos="htb"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-pattern-pep-1"><a href="#fig-pattern-pep-1" aria-hidden="true" tabindex="-1"></a><span class="fu">\resizebox</span>{.65<span class="fu">\textwidth</span>}{!}{<span class="co">%</span></span>
+<div class="sourceCode" id="fig-pattern-pep" data-fig-pos="htb" data-fig-env="figure"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-pattern-pep-1"><a href="#fig-pattern-pep-1" aria-hidden="true" tabindex="-1"></a><span class="fu">\resizebox</span>{.65<span class="fu">\textwidth</span>}{!}{<span class="co">%</span></span>
 <span id="fig-pattern-pep-2"><a href="#fig-pattern-pep-2" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,node distance=12mm]</span>
 <span id="fig-pattern-pep-3"><a href="#fig-pattern-pep-3" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col2}{RGB}{255, 255, 128}</span>
 <span id="fig-pattern-pep-4"><a href="#fig-pattern-pep-4" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col4}{RGB}{240,240,255}</span>
@@ -1249,10 +1249,10 @@ <h4 class="anchored" data-anchor-id="case-study-wildlife-insights">Case Study: W
 <h4 class="anchored" data-anchor-id="pattern-structure-2">Pattern Structure</h4>
 <p>The Distributed Knowledge Pattern comprises specific architectural components designed to enable decentralized data collection, processing, and knowledge sharing. The pattern defines three primary structural elements: autonomous nodes, communication networks, and aggregation mechanisms.</p>
 <p><a href="#fig-pattern_dc" class="quarto-xref">Figure&nbsp;<span>19.5</span></a> illustrates the key components and their interactions within the Distributed Knowledge Pattern. Individual nodes (rectangular shapes) operate autonomously while sharing insights through defined communication channels. The aggregation layer (diamond shape) combines distributed knowledge, which feeds into the analysis layer (oval shape) for processing.</p>
-<div id="fig-pattern_dc" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-env="figure" data-fig-pos="htb">
+<div id="fig-pattern_dc" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-pos="htb" data-fig-env="figure">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-pattern_dc-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<div class="sourceCode" id="fig-pattern_dc" data-fig-env="figure" data-fig-pos="htb"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-pattern_dc-1"><a href="#fig-pattern_dc-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,node distance=24mm]</span>
+<div class="sourceCode" id="fig-pattern_dc" data-fig-pos="htb" data-fig-env="figure"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-pattern_dc-1"><a href="#fig-pattern_dc-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,node distance=24mm]</span>
 <span id="fig-pattern_dc-2"><a href="#fig-pattern_dc-2" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{colorFill1}{RGB}{180,222,240}</span>
 <span id="fig-pattern_dc-3"><a href="#fig-pattern_dc-3" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{colorFill2}{RGB}{219,253,166}</span>
 <span id="fig-pattern_dc-4"><a href="#fig-pattern_dc-4" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{colorLine1}{RGB}{73,89,56}</span>
@@ -1346,10 +1346,10 @@ <h3 data-number="19.6.4" class="anchored" data-anchor-id="adaptive-resource"><sp
 <p>The Adaptive Resource Pattern focuses on enabling systems to dynamically adjust their operations in response to varying resource availability, ensuring efficiency, scalability, and resilience in real-time. This pattern allows systems to allocate resources flexibly depending on factors like computational load, network bandwidth, and storage capacity. The key idea is that systems should be able to scale up or down based on the resources they have access to at any given time.</p>
 <p>Rather than being a standalone pattern, Adaptive Resource Pattern management is often integrated within other system design patterns. It enhances systems by allowing them to perform efficiently even under changing conditions, ensuring that they continue to meet their objectives, regardless of resource fluctuations.</p>
 <p><a href="#fig-patterns_adaptive" class="quarto-xref">Figure&nbsp;<span>19.6</span></a> below illustrates how systems using the Adaptive Resource Pattern adapt to different levels of resource availability. The system adjusts its operations based on the resources available at the time, optimizing its performance accordingly.</p>
-<div id="fig-patterns_adaptive" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-env="figure" data-fig-pos="htb">
+<div id="fig-patterns_adaptive" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-pos="htb" data-fig-env="figure">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-patterns_adaptive-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<div class="sourceCode" id="fig-patterns_adaptive" data-fig-env="figure" data-fig-pos="htb"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-patterns_adaptive-1"><a href="#fig-patterns_adaptive-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,node distance=9mm]</span>
+<div class="sourceCode" id="fig-patterns_adaptive" data-fig-pos="htb" data-fig-env="figure"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-patterns_adaptive-1"><a href="#fig-patterns_adaptive-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,node distance=9mm]</span>
 <span id="fig-patterns_adaptive-2"><a href="#fig-patterns_adaptive-2" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{colorFill1}{RGB}{180,222,240}</span>
 <span id="fig-patterns_adaptive-3"><a href="#fig-patterns_adaptive-3" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{colorFill2}{RGB}{219,253,166}</span>
 <span id="fig-patterns_adaptive-4"><a href="#fig-patterns_adaptive-4" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{colorLine1}{RGB}{73,89,56}</span>

docs/contents/core/ml_systems/ml_systems.html

@@ -1412,10 +1412,10 @@ <h4 class="anchored" data-anchor-id="collaborative-learning">Collaborative Learn
 <h3 data-number="2.6.2" class="anchored" data-anchor-id="real-world-integration"><span class="header-section-number">2.6.2</span> Real-world Integration</h3>
 <p>Design patterns establish a foundation for organizing and optimizing ML workloads across distributed systems. However, the practical application of these patterns often requires combining multiple paradigms into integrated workflows. Thus, in practice, ML systems rarely operate in isolation. Instead, they form interconnected networks where each paradigm—Cloud, Edge, Mobile, and Tiny ML—plays a specific role while communicating with other parts of the system. These interconnected networks follow integration patterns that assign specific roles to Cloud, Edge, Mobile, and Tiny ML systems based on their unique strengths and limitations. Recall that cloud systems excel at training and analytics but require significant infrastructure. Edge systems provide local processing power and reduced latency. Mobile devices offer personal computing capabilities and user interaction. Tiny ML enables intelligence in the smallest devices and sensors.</p>
 <p><a href="#fig-hybrid" class="quarto-xref">Figure&nbsp;<span>2.9</span></a> illustrates these key interactions through specific connection types: “Deploy” paths show how models flow from cloud training to various devices, “Data” and “Results” show information flow from sensors through processing stages, “Analyze” shows how processed information reaches cloud analytics, and “Sync” demonstrates device coordination. Notice how data generally flows upward from sensors through processing layers to cloud analytics, while model deployments flow downward from cloud training to various inference points. The interactions aren’t strictly hierarchical—mobile devices might communicate directly with both cloud services and tiny sensors, while edge systems can assist mobile devices with complex processing tasks.</p>
-<div id="fig-hybrid" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-pos="htb" data-fig-env="figure">
+<div id="fig-hybrid" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-env="figure" data-fig-pos="htb">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-hybrid-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<div class="sourceCode" id="fig-hybrid" data-fig-pos="htb" data-fig-env="figure"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-hybrid-1"><a href="#fig-hybrid-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,line width=0.75pt]</span>
+<div class="sourceCode" id="fig-hybrid" data-fig-env="figure" data-fig-pos="htb"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-hybrid-1"><a href="#fig-hybrid-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,line width=0.75pt]</span>
 <span id="fig-hybrid-2"><a href="#fig-hybrid-2" aria-hidden="true" tabindex="-1"></a><span class="fu">\usetikzlibrary</span>{calc,fit,backgrounds,positioning}</span>
 <span id="fig-hybrid-3"><a href="#fig-hybrid-3" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col2}{RGB}{255, 255, 128}</span>
 <span id="fig-hybrid-4"><a href="#fig-hybrid-4" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col5}{RGB}{170,170,51}</span>
@@ -1517,10 +1517,10 @@ <h3 data-number="2.6.2" class="anchored" data-anchor-id="real-world-integration"
 <h2 data-number="2.7" class="anchored" data-anchor-id="shared-principles"><span class="header-section-number">2.7</span> Shared Principles</h2>
 <p>The design and integration patterns illustrate how ML paradigms—Cloud, Edge, Mobile, and Tiny–interact to address real-world challenges. While each paradigm is tailored to specific roles, their interactions reveal recurring principles that guide effective system design. These shared principles provide a unifying framework for understanding both individual ML paradigms and their hybrid combinations. As we explore these principles, a deeper system design perspective emerges, showing how different ML implementations—optimized for distinct contexts—converge around core concepts. This convergence forms the foundation for systematically understanding ML systems, despite their diversity and breadth.</p>
 <p><a href="#fig-ml-systems-convergence" class="quarto-xref">Figure&nbsp;<span>2.10</span></a> illustrates this convergence, highlighting the relationships that underpin practical system design and implementation. Grasping these principles is invaluable not only for working with individual ML systems but also for developing hybrid solutions that leverage their strengths, mitigate their limitations, and create cohesive, efficient ML workflows.</p>
-<div id="fig-ml-systems-convergence" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-pos="htb" data-fig-env="figure">
+<div id="fig-ml-systems-convergence" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-env="figure" data-fig-pos="htb">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-ml-systems-convergence-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<div class="sourceCode" id="fig-ml-systems-convergence" data-fig-pos="htb" data-fig-env="figure"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-ml-systems-convergence-1"><a href="#fig-ml-systems-convergence-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,line width=0.75pt]</span>
+<div class="sourceCode" id="fig-ml-systems-convergence" data-fig-env="figure" data-fig-pos="htb"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-ml-systems-convergence-1"><a href="#fig-ml-systems-convergence-1" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,line width=0.75pt]</span>
 <span id="fig-ml-systems-convergence-2"><a href="#fig-ml-systems-convergence-2" aria-hidden="true" tabindex="-1"></a><span class="fu">\usetikzlibrary</span>{calc,fit,backgrounds,positioning}</span>
 <span id="fig-ml-systems-convergence-3"><a href="#fig-ml-systems-convergence-3" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col2}{RGB}{255, 255, 128}</span>
 <span id="fig-ml-systems-convergence-4"><a href="#fig-ml-systems-convergence-4" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col5}{RGB}{170,170,51}</span>
@@ -1834,10 +1834,10 @@ <h2 data-number="2.8" class="anchored" data-anchor-id="ml-system-comparison"><sp
 <section id="ml-deployment-decision-framework" class="level2" data-number="2.9">
 <h2 data-number="2.9" class="anchored" data-anchor-id="ml-deployment-decision-framework"><span class="header-section-number">2.9</span> ML Deployment Decision Framework</h2>
 <p>We have examined the diverse paradigms of machine learning systems—Cloud ML, Edge ML, Mobile ML, and Tiny ML—each with its own characteristics, trade-offs, and use cases. Selecting an optimal deployment strategy requires careful consideration of multiple factors.</p>
-<div id="fig-mlsys-playbook-flowchart" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-pos="!t" data-fig-env="figure">
+<div id="fig-mlsys-playbook-flowchart" class="quarto-float quarto-figure quarto-figure-center anchored" data-fig-env="figure" data-fig-pos="!t">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-mlsys-playbook-flowchart-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<div class="sourceCode" id="fig-mlsys-playbook-flowchart" data-fig-pos="!t" data-fig-env="figure"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-mlsys-playbook-flowchart-1"><a href="#fig-mlsys-playbook-flowchart-1" aria-hidden="true" tabindex="-1"></a><span class="fu">\resizebox</span>{.65<span class="fu">\textwidth</span>}{!}{<span class="co">%</span></span>
+<div class="sourceCode" id="fig-mlsys-playbook-flowchart" data-fig-env="figure" data-fig-pos="!t"><pre class="sourceCode tikz code-with-copy"><code class="sourceCode latex"><span id="fig-mlsys-playbook-flowchart-1"><a href="#fig-mlsys-playbook-flowchart-1" aria-hidden="true" tabindex="-1"></a><span class="fu">\resizebox</span>{.65<span class="fu">\textwidth</span>}{!}{<span class="co">%</span></span>
 <span id="fig-mlsys-playbook-flowchart-2"><a href="#fig-mlsys-playbook-flowchart-2" aria-hidden="true" tabindex="-1"></a><span class="kw">\begin</span>{<span class="ex">tikzpicture</span>}[font=<span class="fu">\small\sf</span>,line width=0.75pt]</span>
 <span id="fig-mlsys-playbook-flowchart-3"><a href="#fig-mlsys-playbook-flowchart-3" aria-hidden="true" tabindex="-1"></a><span class="fu">\usetikzlibrary</span>{calc,fit,backgrounds,positioning}</span>
 <span id="fig-mlsys-playbook-flowchart-4"><a href="#fig-mlsys-playbook-flowchart-4" aria-hidden="true" tabindex="-1"></a><span class="fu">\definecolor</span>{col2}{RGB}{255,255,128}</span>

docs/contents/core/references.html

@@ -3091,7 +3091,7 @@ <h1 class="title">References</h1>
 </div>
 <div id="ref-rachwan2022winning" class="csl-entry" role="listitem">
 Rachwan, John, Daniel Zügner, Bertrand Charpentier, Simon Geisler,
-Morgane Ayle, and Stephan Günnemann. 2022. <span>“Winning the Lottery
+Morgane Ayle, and Stephan Günnemann. 2022. <span>���Winning the Lottery
 Ahead of Time: Efficient Early Network Pruning.”</span> In
 <em>International Conference on Machine Learning</em>, 18293–309. PMLR.
 </div>