EE · Fall 2026 · Undergraduate

Deep Learning

From neural network foundations to transformers, generative models, and deep reinforcement learning.

Course Overview

This course develops a practical and conceptual understanding of modern deep learning. We begin with the mathematical and computational foundations of neural networks, then study the architectures that power current vision, language, graph, and generative AI systems. The final part of the course introduces deep reinforcement learning.

Instructor
Prof. Nuri Kim
Term
Fall 2026
Meetings
Mon & Wed
09:00–10:15
Classroom
Engineering Building 7
Room 114

Logistics

Learning goals

  • Explain the principles behind training and regularizing deep neural networks.
  • Implement and evaluate modern architectures with PyTorch.
  • Connect model design choices to vision, language, graph, and generative tasks.
  • Read and discuss current deep learning research critically.

Recommended background

  • Comfort with Python programming.
  • Introductory linear algebra, calculus, and probability.
  • Basic machine learning concepts are helpful but will be reviewed.

Assessment details and exact deadlines will be announced through the JBNU LMS before the semester begins.

Lectures

Part I
Foundations
00Course IntroductionTBA
01Foundations of Deep LearningTBA
02OptimizationTBA
03RegularizationTBA
04Deep Feedforward NetworksTBA
Part II
Architectures
05Convolutional Neural NetworksTBA
06Recurrent Neural NetworksTBA
07AttentionTBA
08TransformersTBA
09Vision ArchitecturesTBA
10Transformers in NLPTBA
11Graph Neural NetworksTBA
12Additional Deep ModelsTBA
Part III
Generative AI
13Introduction to Deep Generative ModelsTBA
14Score-based Diffusion Models ITBA
15Score-based Diffusion Models IITBA
16Score-based Diffusion Models IIITBA
17Score-based Diffusion Models IVTBA
18Variational AutoencodersTBA
19Generative Adversarial NetworksTBA
Part IV
Decision Making
20Deep Reinforcement Learning ITBA
21Deep Reinforcement Learning IITBA
22Deep Reinforcement Learning IIITBA

The lecture order is tentative and may change as the course progresses. Slides and readings will be linked here when released.

Assignments

Assignment 1

Modern Architectures

CNNs, transformers from scratch, vision transformers, and BERT fine-tuning.

Assignment 2

Generative Models

Diffusion-based generation and practical text-to-image modeling.

Term Project

Open-ended Deep Learning

A multi-stage project connecting model implementation, evaluation, and communication.

Programming Projects
20%
Midterm Exam
25%
Final Exam
25%
Final Project
25%
Attendance
5%+

Assignment files, submission instructions, collaboration rules, and deadlines are available only to enrolled students through the JBNU LMS.

Resources

  • Course platformJBNU LMS
  • Primary frameworkPyTorch
  • ReferenceGoodfellow, Bengio, and Courville, Deep Learning
  • ReferenceSutton and Barto, Reinforcement Learning: An Introduction
  • Contactnuri.kim@jbnu.ac.kr