1<!DOCTYPE html> 2<html lang="en"> 3<head> 4 <meta charset="utf-8"> 5 <meta name="viewport" content="width=device-width, initial-scale=1"> 6 <meta name="description" content="DDA4220: Deep Learning and Applications, Fall 2026, taught by Zhen Liu at CUHK-Shenzhen. Course schedule, materials, assessment, and teaching staff."> 7 <meta name="theme-color" content="#245a9b"> 8 <title>DDA4220 · Deep Learning and Applications · Fall 2026</title> 9 <link rel="icon" href="favicon.svg" type="image/svg+xml"> 10 <link rel="stylesheet" href="course.css"> 11
11<script src="course.js" defer></script>
11 12</head> 13<body> 14 <a class="skip-link" href="#main">Skip to course content</a> 15 <header class="masthead"> 16 <div class="masthead-inner"> 17 <a class="course-mark" href="../../index.html">Zhen Liu <span class="mark-divider" aria-hidden="true">/</span> Teaching</a> 18 <span class="institution">CUHK-Shenzhen <span class="institution-divider" aria-hidden="true">·</span> School of Data Science</span> 19 </div> 20 </header> 21 22 <main id="main"> 23 <section class="hero page-width" id="top" aria-labelledby="course-title"> 24 <div class="hero-copy"> 25 <p class="eyebrow">DDA4220 <span aria-hidden="true">·</span> Fall 2026</p> 26 <h1 id="course-title">Deep Learning and Applications</h1> 27 </div> 28 </section> 29 30 <div class="course-strip"> 31 <dl class="page-width quick-facts"> 32 <div><dt>Lectures</dt><dd>Tue & Thu · 10:30â12:00</dd></div> 33 <div><dt>Room</dt><dd>J_E205</dd></div> 34 <div><dt>First lecture</dt><dd><time datetime="2026-09-08">Tuesday, September 8</time></dd></div> 35 </dl> 36 </div> 37 38 <div class="page-width course-layout"> 39 <aside class="section-index"> 40 <nav aria-label="Course sections"> 41 <a class="is-active" href="#overview" aria-current="location">Overview</a> 42 <a href="#schedule">Schedule</a> 43 <a href="#materials">Reading</a> 44 <a href="#information">Assessment & policies</a> 45 <a href="#staff">Teaching team</a> 46 </nav> 47 </aside> 48 49 <div class="course-content"> 50 <section class="content-section overview-section" id="overview" aria-labelledby="overview-title"> 51 <div class="section-heading"><h2 id="overview-title">About the course</h2></div> 52 <p>This course introduces the foundations of deep learning: neural network architectures, training objectives, optimization, regularization, and the engineering practices behind effective models. We explore supervised, self-supervised, unsupervised, and reinforcement learning, with applications in computer vision, natural language processing, and robotics.</p> 53 <p class="overview-followup">You will learn to design, implement, and train models with tools such as PyTorch and TensorFlow, gaining practical experience through hands-on assignments and projects. Advanced topics will be introduced as time permits.</p> 54 <details class="prerequisites"><summary>Prerequisites & preparation <span aria-hidden="true">+</span></summary><div class="details-body"><p><strong>Python proficiency is required.</strong> The course prerequisites are:</p><ul class="prerequisite-codes"><li>CSC1001 / CSC1003</li><li>MAT2040 / MAT2041, or equivalent courses</li><li>DDA3020 / CSC4020 / DDA2020 / FTE4560</li></ul><p>Prior coursework or projects in computer vision or natural language processing, and familiarity with PyTorch or TensorFlow, are helpful. There are no co-requisites.</p></div></details> 55 </section> 56 57 <section class="content-section" id="schedule" aria-labelledby="schedule-title"> 58 <div class="section-heading"><h2 id="schedule-title">Lecture schedule</h2></div> 59 <p class="schedule-note">The course schedule is tentative and subject to change. Lecture materials will be linked below as they become available.</p> 60 <div class="schedule-scroll" role="region" aria-label="Lecture schedule and deadlines" tabindex="0"> 61 <table class="schedule-table"> 62 <caption class="visually-hidden">Dated lecture plan for Fall 2026. October 1 and 6 have no class during the October 1â7 National Day holiday. The midterm is listed separately. Deadlines appear alongside lectures with their actual due dates stated explicitly. A dash in Materials means materials have not been posted, or are not applicable for holidays.</caption> 63 <thead><tr><th scope="col" class="lecture-column">Lecture</th><th scope="col" class="date-column">Date</th><th scope="col">Topics</th><th scope="col" class="schedule-status materials-column">Materials</th><th scope="col" class="deadlines-column">Deadlines</th></tr></thead> 64 <tbody> 65 <tr><th scope="row">01</th><td class="schedule-date"><time datetime="2026-09-08" aria-label="Tuesday, September 8, 2026">
65Sep 8</time></td><td><h3>Course introduction & history</h3></td><td class="schedule-status"><a href="materials/lectures/L01.pdf" aria-label="Lecture 1 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 66 <tr><th scope="row">02</th><td class="schedule-date"><time datetime="2026-09-10" aria-label="Thursday, September 10, 2026">Sep 10</time></td><td><h3>Neural networks: representation and learning</h3></td><td class="schedule-status"><a href="materials/lectures/L02.pdf" aria-label="Lecture 2 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 67 <tr><th scope="row">03</th><td class="schedule-date"><time datetime="2026-09-15" aria-label="Tuesday, September 15, 2026">Sep 15</time></td><td><h3>Training neural networks I: backpropagation and optimizers</h3></td><td class="schedule-status"><a href="materials/lectures/L03.pdf" aria-label="Lecture 3 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 68 <tr><th scope="row">04</th><td class="schedule-date"><time datetime="2026-09-17" aria-label="Thursday, September 17, 2026">Sep 17</time></td><td><h3>Training neural networks II: initialization, normalization</h3></td><td class="schedule-status"><a href="materials/lectures/L04.pdf" aria-label="Lecture 4 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 69 <tr><th scope="row">05</th><td class="schedule-date"><time datetime="2026-09-22" aria-label="Tuesday, September 22, 2026">Sep 22</time></td><td><h3>Training neural networks III: generalization, regularization, and training practice</h3></td><td class="schedule-status"><a href="materials/lectures/L05.pdf" aria-label="Lecture 5 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 70 <tr><th scope="row">06</th><td class="schedule-date"><time datetime="2026-09-24" aria-label="Thursday, September 24, 2026">Sep 24</time></td><td><h3>Convolutional networks and object recognition</h3></td><td class="schedule-status"><a href="materials/lectures/L06.pdf" aria-label="Lecture 6 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 71 <tr><th scope="row">07</th><td class="schedule-date"><time datetime="2026-09-29" aria-label="Tuesday, September 29, 2026">Sep 29</time></td><td><h3>Visual representations and transfer learning</h3></td><td class="schedule-status"><a href="materials/lectures/L07.pdf" aria-label="Lecture 7 slides (PDF)">Slides (PDF)</a></td><td class="schedule-deadlines"></td></tr> 72 <tr class="holiday-row"><th scope="row"><span class="visually-hidden">No lecture</span>â</th><td class="schedule-date"><time datetime="2026-10-01" aria-label="Thursday, October 1, 2026">Oct 1</time></td><td><h3>National holiday â no class</h3></td><td class="schedule-status"><span aria-label="Not applicable">â</span></td><td class="schedule-deadlines"></td></tr> 73 <tr class="holiday-row"><th scope="row"><span class="visually-hidden">No lecture</span>â</th><td class="schedule-date"><time datetime="2026-10-06" aria-label="Tuesday, October 6, 2026">Oct 6</time></td><td><h3>National holiday â no class</h3></td><td class="schedule-status"><span aria-label="Not applicable">â</span></td><td class="schedule-deadlines"></td></tr> 74 <tr><th scope="row">08</th><td class="schedule-date"><time datetime="2026-10-08" aria-label="Thursday, October 8, 2026">Oct 8</time></td><td><h3>Dense prediction and segmentation</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 75 <tr><th scope="row">09</th><td class="schedule-date"><time datetime="2026-10-13" aria-label="Tuesday, October 13, 2026">Oct 13</time></td><td><h3>Object detection and instance segmentation</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 76 <tr><th scope="row">10</th><td class="schedule-date"><time datetime="2026-10-15" aria-label="Thursday, October 15, 2026">Oct 15</time></td><td><h3>Sequence models and language modeling</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"><a href="materials/assignments/A1_Student_Package.zip" download>Assignment 1 (due <time datetime="2026-10-16">Oct 16</time>)</a><br><a href="project.html#proposal">Proposal (due <time datetime="2026-10-16">Oct 16</time>)</a></td></tr> 77 <tr><th scope="row">11</th><td class="schedule-date"><time datetime="2026-10-20" aria-label="Tuesday, October 20, 2026">Oct 20</time></td><td><h3>Sequence-to-sequence models and attention</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 78 <tr><th scope="row">12</th><td class="schedule-date"><time datetime="2026-10-22" aria-label="Thursday, October 22, 2026">Oct 22</time></td><td><h3>Transformers and Vision Transformers</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 79 <tr><th scope="row">13</th><td class="schedule-date"><time datetime="2026-10-27" aria-label="Tuesday, October 27, 2026">Oct 27</time></td><td><h3>Self-supervised learning</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 80 <tr><th scope="row">14</th><td class="schedule-date"><time datetime="2026-10-29" aria-label="Thursday, October 29, 2026">Oct 29</time></td><td><h3>Autoencoders & variational autoencoders</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 81 <tr><th scope="row">15</th><td class="schedule-date"><time datetime="2026-11-03" aria-label="Tuesday, November 3, 2026">Nov 3</time></td><td><h3>Midterm review</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 82 <tr class="exam-row" id="midterm-exam"><th scope="row"><span class="visually-hidden">Assessment</span>â</th><td class="schedule-date"><time datetime="2026-11-05" aria-label="Thursday, November 5, 2026">Nov 5</time></td><td><h3>Midterm exam (tentative)</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 83 <tr><th scope="row">16</th><td class="schedule-date"><time datetime="2026-11-10" aria-label="Tuesday, November 10, 2026">Nov 10</time></td><td><h3>Generative adversarial networks</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 84 <tr><th scope="row">17</th><td class="schedule-date"><time datetime="2026-11-12" aria-label="Thursday, November 12, 2026">Nov 12</time></td><td><h3>Diffusion models</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 85 <tr><th scope="row">18</th><td class="schedule-date"><time datetime="2026-11-17" aria-label="Tuesday, November 17, 2026">Nov 17</time></td><td><h3>Conditional generation and flow matching</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"><a href="project.html#midterm-report">Midterm report (due <time datetime="2026-11-18">Nov 18</time>)</a></td></tr> 86 <tr><th scope="row">19</th><td class="schedule-date"><time datetime="2026-11-19" aria-label="Thursday, November 19, 2026">Nov 19</time></td><td><h3>Reinforcement Learning: Foundations and Value-Based Methods</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 87 <tr><th scope="row">20</th><td class="schedule-date"><time datetime="2026-11-24" aria-label="Tuesday, November 24, 2026">Nov 24</time></td><td><h3>Reinforcement Learning: Policy Gradients and ActorâCritic Methods</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 88 <tr><th scope="row">21</th><td class="schedule-date"><time datetime="2026-11-26" aria-label="Thursday, November 26, 2026">Nov 26</time></td><td><h3>Language foundation models</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 89 <tr><th scope="row">22</th><td class="schedule-date"><time datetime="2026-12-01" aria-label="Tuesday, December 1, 2026">Dec 1</time></td><td><h3>Multimodal & generative foundation models</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 90 <tr><th scope="row">23</th><td class="schedule-date"><time datetime="2026-12-03" aria-label="Thursday, December 3, 2026">Dec 3</time></td><td><h3>3D vision</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 91 <tr><th scope="row">24</th><td class="schedule-date"><time datetime="2026-12-08" aria-label="Tuesday, December 8, 2026">Dec 8</time></td><td><h3>Robot learning</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 92 <tr><th scope="row">25</th><td class="schedule-date"><time datetime="2026-12-10" aria-label="Thursday, December 10, 2026">Dec 10</time></td><td><h3>Advanced topics I</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 93 <tr><th scope="row">26</th><td class="schedule-date"><time datetime="2026-12-15" aria-label="Tuesday, December 15, 2026">Dec 15</time></td><td><h3>Advanced topics II</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 94 <tr><th scope="row">27</th><td class="schedule-date"><time datetime="2026-12-17" aria-label="Thursday, December 17, 2026">Dec 17</time></td><td><h3>Course review & discussion</h3></td><td class="schedule-status"><span aria-label="Not yet posted">â</span></td><td class="schedule-deadlines"></td></tr> 95 </tbody> 96 </table> 97 </div> 98 </section> 99 100 <section class="content-section" id="materials" aria-labelledby="materials-title"> 101 <div class="section-heading"><h2 id="materials-title">Recommended reading</h2></div> 102 <div class="reading-list"> 103 <p>No textbook is required.</p> 104 <a class="book-link" href="https://www.deeplearningbook.org/"><div><h3>Deep Learning</h3><p>Ian Goodfellow, Yoshua Bengio & Aaron Courville<span>MIT Press</span></p></div></a> 105 <a class="book-link" href="https://www.bishopbook.com/"><div><h3>Deep Learning: Foundations and Concepts</h3><p>Christopher M. Bishop & Hugh Bishop<span>Springer</span></p></div></a> 106 </div> 107 </section> 108 109 <section class="content-section" id="information" aria-labelledby="information-title"> 110 <div class="section-heading"><h2 id="information-title">Assessment & policies</h2></div> 111 <dl class="assessment-list"> 112 <div><dt>3 homework assignments</dt><dd>45%</dd></div> 113 <div><dt><a href="project.html">Final project</a><span>Project description, milestones, and requirements.</span></dt><dd>25%</dd></div> 114 <div><dt><a href="#midterm-exam">Midterm exam</a></dt><dd>30%</dd></div> 115 </dl> 116 <div class="assignment-section" id="assignments" aria-labelledby="assignments-title"> 117 <h3 id="assignments-title">Assignments</h3> 118 <ul class="assignment-list"> 119 <li><span class="assignment-details"><span class="assignment-title">Assignment 1</span><a class="assignment-download" href="materials/assignments/A1_Student_Package.zip" aria-label="Download Assignment 1 package (ZIP)" download>Package (ZIP)</a></span><span class="assignment-status">Due <time datetime="2026-10-16">Oct 16</time></span></li> 120 <li><span class="assignment-title">
120Assignment 2</span><span class="assignment-status">Not yet released</span></li> 121 <li><span class="assignment-title">Assignment 3</span><span class="assignment-status">Not yet released</span></li> 122 </ul> 123 </div> 124 <div class="policy-note"><h3>Course policies</h3><p>See <a href="materials/lectures/L01.pdf#page=85">L01</a> for coursework and AI-use guidance.</p><p><strong>Late submissions:</strong> You have <strong>3 grace days for all homework combined</strong>. Each 24-hour period, or part of one, uses 1 grace day: 23 hours late uses 1 day; 25 hours uses 2.</p></div> 125 </section> 126 127 <section class="content-section staff-section" id="staff" aria-labelledby="staff-title"> 128 <div class="section-heading"><h2 id="staff-title">Teaching team & office hours</h2></div> 129 <div class="staff-grid"> 130 <div class="instructor"><div><p class="staff-role">Instructor</p><h3>Zhen Liu</h3><span class="staff-email">[first name + last name]@cuhk.edu.cn</span><a class="plain-link" href="../../index.html">Personal webpage</a></div></div> 131 <div class="staff-details" id="office-hours"><h3>Instructor office hours</h3><p class="office-time">Tuesday 09:00â10:00 · every other week</p><p>First session: <time datetime="2026-09-15">Tuesday, September 15, 2026</time>.</p><p>Daoyuan 323B (<span lang="zh-Hans">éè¿æ¥¼323B</span>)</p><p class="office-policy">Held only by request; please email at least one day in advance.</p></div> 132 </div> 133 <div class="ta-list" aria-label="Teaching assistants"> 134 <div class="ta-person"> 135 <p class="staff-role">Lead teaching assistant</p> 136 <h3>Wenqian Zhang <span lang="zh-Hans">å¼ æè°¦</span></h3> 137 <a class="staff-email" href="mailto:[email protected]">[email protected]</a> 138 <div class="ta-office-hours"> 139 <h4>Office hours</h4> 140 <p>Tuesday 17:00â18:00<br>Thursday 14:30â15:30</p> 141 <p class="ta-location">Zhixin (<span lang="zh-Hans">ç¥æ°æ¥¼</span>) 408</p> 142 </div> 143 </div> 144 <div class="ta-person"> 145 <p class="staff-role">Teaching assistant</p> 146 <h3>Ruixiang Wang <span lang="zh-Hans">ççç¿</span></h3> 147 <a class="staff-email" href="mailto:[email protected]">[email protected]</a> 148 <div class="ta-office-hours"> 149 <h4>Office hours</h4> 150 <p>Saturday 17:00â18:00<br>Sunday 17:00â18:00</p> 151 <p class="ta-location">Teaching Complex B 609 (<span lang="zh-Hans">æå¦ç»¼å楼B 609</span>)</p> 152 </div> 153 </div> 154 <div class="ta-person"> 155 <p class="staff-role">Teaching assistant</p> 156 <h3>Yifan Peng <span lang="zh-Hans">å½ä¸å¡</span></h3> 157 <a class="staff-email" href="mailto:[email protected]">[email protected]</a> 158 <div class="ta-office-hours"> 159 <h4>Office hours</h4> 160 <p>Monday 19:00â20:00<br>Wednesday 19:00â20:00</p> 161 <p class="ta-location">Zhixin (<span lang="zh-Hans">ç¥æ°æ¥¼</span>) 429</p> 162 </div> 163 </div> 164 </div> 165 </section> 166 <section class="content-section acknowledgments" id="acknowledgments" aria-labelledby="acknowledgments-title"> 167 <div class="section-heading"><h2 id="acknowledgments-title">Acknowledgments</h2></div> 168 <p>This course draws inspiration from <a href="https://cs231n.stanford.edu/">Stanford CS231n</a>, <a href="https://sites.cc.gatech.edu/classes/AY2025/cs7643_fall/">Georgia Tech CS 4644/7643 (Danfei Xu)</a>, <a href="https://deeplearning.cs.cmu.edu/F25/index.html">CMU 11-785</a>, and <a href="https://introtodeeplearning.com/">MIT 6.S191</a>. We also borrow a few figures from their teaching materials.</p> 169 </section> 170 </div> 171 </div> 172 </main> 173 174 <footer class="page-width footer"> 175 <a class="footer-course" href="#top">DDA4220 <span>Deep Learning and Applications</span></a> 176 <span>Fall 2026 · CUHK-Shenzhen</span> 177 <a class="back-top" href="#top">Back to top <span aria-hidden="true">â</span></a> 178 </footer> 179</body> 180</html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.