statistical-ml.html

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Era Details: Statistical Machine Learning</title>
    <link rel="stylesheet" href="style.css">
    <style>
        .details-body { max-width: 800px; margin: 0 auto; padding: 40px 20px; text-align: left; }
        .hero-section { border-bottom: 2px solid var(--primary); padding-bottom: 20px; margin-bottom: 40px; }
        .quote { font-style: italic; font-size: 1.5rem; color: #38bdf8; margin: 10px 0; }
        .back-btn { display: inline-block; margin-bottom: 20px; color: var(--primary); text-decoration: none; font-weight: bold; }
        
        .sub-timeline { border-left: 2px solid #334155; padding-left: 20px; margin-left: 10px; }
        .milestone { margin-bottom: 30px; position: relative; }
        .milestone::before { content: ""; position: absolute; left: -26px; top: 5px; width: 10px; height: 10px; background: #38bdf8; border-radius: 50%; }
        
        .methodology-box { background: #1e293b; border-left: 4px solid var(--primary); padding: 20px; margin: 20px 0; }
        .tech-list { display: grid; grid-template-columns: repeat(auto-fit, minmax(200px, 1fr)); gap: 15px; }
        .tech-tag { background: #0f172a; border: 1px solid #334155; padding: 10px; border-radius: 5px; font-family: monospace; font-size: 0.85rem; }
    </style>
</head>
<body>
    <div class="details-body">
        <a href="index.html" class="back-btn">← Back to Timeline</a>
        
        <div class="hero-section">
            <span class="year" style="color: #38bdf8;">1995 – 2010</span>
            <h1>Statistical Machine Learning</h1>
            <p class="quote">"Don't tell the computer what to do; show it what you've done."</p>
        </div>

        <section>
            <h2>Chronology of the Probabilistic Shift</h2>
            <div class="sub-timeline">
                <div class="milestone">
                    <h4>1995: Support Vector Machines (SVM) Popularity</h4>
                    <p>Cortes and Vapnik publish their work on SVMs, providing a powerful mathematical framework for classification that outperformed early neural nets.</p>
                </div>
                <div class="milestone">
                    <h4>1997: Deep Blue vs. Garry Kasparov</h4>
                    <p>IBM's Deep Blue defeats the world chess champion. While largely "brute-force," it utilized sophisticated evaluation functions learned from grandmaster games.</p>
                </div>
                <div class="milestone">
                    <h4>2001: Random Forests Algorithm</h4>
                    <p>Leo Breiman introduces Random Forests, showing that an ensemble of many "weak" decision trees could create a very "strong" and stable predictor.</p>
                </div>
                <div class="milestone">
                    <h4>2006: The Netflix Prize</h4>
                    <p>Netflix offers $1M to improve their recommendation engine, catalyzing massive research into Collaborative Filtering and Matrix Factorization.</p>
                </div>
                <div class="milestone">
                    <h4>2009: ImageNet Launch</h4>
                    <p>Fei-Fei Li launches ImageNet, a massive labeled dataset of 14 million images, creating the "competition" that would eventually trigger the Deep Learning era.</p>
                </div>
            </div>
        </section>

        <section style="margin-top: 40px;">
          <h2>The "Scientific" Methodologies of ML</h2>
          <p>This era moved AI from "hacking" to a structured engineering discipline. These four pillars are what made Statistical ML reliable:</p>
          
          <div class="methodology-box">
              <h4>1. Cross-Validation (The Gold Standard)</h4>
              <p>To ensure a model didn't just "memorize" data, engineers divided datasets into <strong>Training, Validation, and Test sets</strong>. Techniques like <em>K-Fold Cross-Validation</em> allowed models to be tested on multiple subsets of data to prove their stability.</p>
          </div>
      
          <div class="methodology-box">
              <h4>2. Regularization (L1 & L2)</h4>
              <p>Methodologies like <strong>Lasso (L1)</strong> and <strong>Ridge (L2)</strong> regression were introduced to prevent "Overfitting." They added a mathematical penalty for complexity, forcing the model to stay simple and generalize better to the real world.</p>
          </div>
      
          <div class="methodology-box">
              <h4>3. Ensemble Learning</h4>
              <p>The methodology of combining multiple models to get one superior result. 
                  <ul>
                      <li><strong>Bagging:</strong> Training models in parallel (e.g., Random Forests).</li>
                      <li><strong>Boosting:</strong> Training models in sequence, where each new model fixes the errors of the previous one (e.g., AdaBoost, XGBoost).</li>
                  </ul>
              </p>
          </div>
      
          <div class="methodology-box">
              <h4>4. Dimensionality Reduction</h4>
              <p>As data grew, models became overwhelmed. Methodologies like <strong>PCA (Principal Component Analysis)</strong> and <strong>LDA</strong> were used to "compress" hundreds of variables into a few key components without losing the essential information.</p>
          </div>
      </section>
      
      <section style="margin-top: 40px;">
          <h2>The Paradigms of Learning</h2>
          <div class="tech-grid">
              <div class="tech-item">
                  <h4>Supervised Learning</h4>
                  <p>Learning with a teacher. The model is given inputs and the correct answers (labels). Goal: Predict the label for new data.</p>
              </div>
              <div class="tech-item">
                  <h4>Unsupervised Learning</h4>
                  <p>Learning without labels. The model looks for hidden structures or clusters in the data (e.g., grouping customers by behavior).</p>
              </div>
              <div class="tech-item">
                  <h4>Reinforcement Learning (Early)</h4>
                  <p>Learning through trial and error. Agents receive "rewards" or "penalties" to learn a policy (e.g., TD-Learning used in early game AI).</p>
              </div>
          </div>
      </section>
        <section style="margin-top: 40px;">
            <h2>Key Concepts & Breakthroughs</h2>
            <ul>
                <li><strong>The Curse of Dimensionality:</strong> Learning how to handle data with thousands of features without the model becoming "lost" in the noise.</li>
                <li><strong>Feature Engineering:</strong> The manual process of selecting which variables (e.g., the frequency of the word "Free" in an email) are important for the model.</li>
                <li><strong>Overfitting vs. Underfitting:</strong> The struggle to make a model that performs well on *new* data, not just the data it was trained on.</li>
                <li><strong>The Rise of Big Data:</strong> The realization that more data often beats a better algorithm (The "Unreasonable Effectiveness of Data").</li>
            </ul>
        </section>
    </div>
</body>
</html>