Skip to content

Schedule (Spring 2025)

Lectures

Disclaimer: All lecture materials are available on the course GitHub repository. The majority of the materials are borrowed from other (fantastic) courses, specifically, from CMU 11-785, Stanford CS231n, and Stanford CS224n.

Fundamentals

  • Introduction to CSC 561 and Course Logistics
  • Perceptron
  • MLPs: Representation
  • MLPs: Learning, Empirical Risk Minimization
  • Gradient Descent, Backpropagation
  • Vectorized Backpropagation, Computational Graphs
  • Optimization (Second Order Methods, RProp, Momentum)
  • Optimization (Mini-Batches, RMSProp, Adam)
  • Regularization, Activation Functions, Weights Initialization, Dropout, Data Augmentation, Hyperparameter Tuning

Vision

  • Image Filtering, Convolutions
  • Convolutional Neural Networks (CNNs), Batch Normalization
  • CNN Architectures (AlexNet, VGG, ResNet)
  • Image Segmentation, Transposed Convolutions
  • Object Detection (R-CNN, Fast R-CNN, Faster R-CNN, YOLO)
  • Vision Transformers (ViT)

Language

  • Word embeddings, Word2Vec, n-Gram Language Models
  • RNNs, LSTMs
  • Encoder/Decoder architectures, Machine Translation, Attention
  • Self-Attention, Transformers (encoders and decoders)
  • Transformers (BERT, T5, GPT), Pre-training, and Fine-tuning

Assignments

  • Quiz 1 (Understanding deep learning requires rethinking generalization), date 02/18
  • Assignment 1 (MLPs, Batch Gradient Descent), due 03/05 11:59pm
  • Quiz 2 (A ConvNet for the 2020s), date 03/27
  • Assignment 2 (Image Classification with CNNs, Image Caption Generation with RNNs), due 04/18 11:59pm
  • Quiz 3 (Masked Autoencoders are Scalable Vision Learners), date 04/29

Project

  • Deliverables, due 05/05 11:59pm via Gradescope
    • Final Report (PDF)
    • Presentation - Slide deck (PDF)
  • Presentations (15 mins each team), 05/06 2p at TBA